DETC2001/DAC EFFECTIVE GENERATION OF PARETO SETS USING GENETIC PROGRAMMING

Size: px
Start display at page:

Download "DETC2001/DAC EFFECTIVE GENERATION OF PARETO SETS USING GENETIC PROGRAMMING"

Transcription

1 Proceedings of DETC 0 ASME 00 Design Engineering Technical Conferences and Computers and Information in Engineering Conference Pittsburgh, PA, September 9-, 00 DETC00/DAC-094 EFFECTIVE GENERATION OF PARETO SETS USING GENETIC PROGRAMMING John Eddy Graduate Research Assistant Department of Mechanical and Aerospace Engineering University at Buffalo johneddy@eng.buffalo.edu Kemper Lewis Assistant Professor Department of Mechanical and Aerospace Engineering Corresponding Author University at Buffalo kelewis@eng.buffalo.edu ABSTRACT Many designers concede that there is typically more than one measure of performance for an artifact. Often, a large system is decomposed into smaller subsystems each having its own set of objectives, constraints, and parameters. The performance of the final design is a function of the performances of the individual subsystems. It then becomes necessary to consider the tradeoffs that occur in a multi-objective design problem. The complete solution to a multi-objective optimization problem is the entire set of non-dominated configurations commonly referred to as the Pareto set. Common methods of generating points along a Pareto frontier involve repeated conversion of multi-objective problems into single objective problems using weights. These methods have been shown to perform poorly when attempting to populate a Pareto frontier. This work presents an efficient means of generating a thorough spread of points along a Pareto frontier using genetic programming. KEYWORDS Genetic Algorithms, Heuristic Optimization, Multi-Objective Optimization. MOGA, Pareto Frontiers INTRODUCTION Engineering design problems commonly require consideration of more than one measure of performance or objectives. The objectives in a multi-objective design problem may relate differently to one another. Two objectives may be in competition, meaning that improvement of one typically comes at the expense of the other. Tradeoffs must be considered when exploring a problem of this nature. For example, it may be desirable in a design scenario to reduce two quantities simultaneously (ex. cost and weight). However, a unit reduction in weight may necessitate an increase in cost and the designer must then decide if he/she is willing to accept the tradeoff of increased cost for a reduction in weight. Objectives may also be in cooperation with each other, meaning that improvement of one typically accompanies improvement of the other. In this case, there is a single superior design and tradeoffs do not occur. The third possibility is that there is no relationship, as is the case if the objectives have no variables in common. The later two cases are less interesting than the first and are therefore not considered further in this paper. When multiple competing objectives exist, the optimum is no longer a design point but an entire set of non-dominated design points. This set is commonly referred to as the Pareto set []. The Pareto set is composed of Pareto optimal solutions. A feasible design variable vector, x, is Pareto optimal if and only if there is no feasible design variable vector, x, with the characteristics, f i ( x) f i ( x ) for all i, i =, n () fi ( x ) < fi( x ) for at least one i, n where n is the number of objectives. Copyright 00 by ASME

2 The primary issue addressed in this paper is how to improve design concepts in a multiobjective framework. Specifically, how can a Pareto frontier be efficiently and thoroughly populated? Common methods of generating points along a Pareto frontier involve repeated conversion of multi-objective problems into single objective problems. These methods have been shown to perform poorly when attempting to populate a Pareto frontier for many types of problems. Common pitfalls of weighted sum methods include: inability to generate a uniform sampling of a frontier [, 3]; inability to generate points in non-convex portions of a frontier [4, 5]; a non-intuitive relationship between combinatorial parameters (weights, etc.) and performances; and poor efficiency (can require an excessive number of function evaluations). A great deal of effort has been put into overcoming these pitfalls. The approach taken in this paper for solving multiobjective problems involves the use of genetic programming. We present a multiobjective genetic algorithm (MOGA) tailored such that the entire Pareto frontier is sought in a single optimization run without repeated conversion from a multiobjective to single objective problem. Having presented a brief discussion of the nature of multiobjective optimization problems, Section. presents some background on means of solving these problems, Section. provides some detailed information regarding the specific genetic algorithm used for this work, and Section includes two case studies to demonstrate the effectiveness of the multiobjective genetic algorithm developed in this work. TECHNICAL BACKGROUND. Multi-Objective Solution Techniques Because weighted sum methods have difficulty in finding and generating Pareto frontiers, many researchers have turned to other methods to generate Pareto frontiers. A straightforward approach is to apply a grid search algorithm, as employed to develop Pareto frontiers in vehicle dynamics simulation problems [6]. The Normal Boundary Intersection (NBI) method [7] also overcomes many difficulties associated with the weighted sums. The NBI generates evenly spaced Pareto points for an even spread of weights, and the spacing of the points is independent of the relative scaling of the objectives. NBI is not limited to bi-objective problems, and is primarily used for tracing the Pareto frontier. Many other researchers have employed other heuristic techniques that do not require conversion to single objectives. Messac and Sundararaj [8] employ Physical Programming to generate a distribution of points along the Pareto frontier. Physical Programming is an optimization method that does not rely on weights, but uses designer preferences in the form of metric classes in the optimization process. The Interactive Sequential Hybrid Optimization Technique (I-SHOT) developed by Narayanan and Azarm [9] is another technique for finding Pareto points. It allows the user to interact at each iteration to help specialize the solution process. It is carried out by repeated application of a simple genetic algorithm. The advantage of this method is that each iteration generates a Pareto solution that is not overly close to a previous solution. The disadvantage of this type of method is that it typically requires a huge number of function evaluations to populate the Pareto Frontier. Some of the tradeoffs in multiobjective optimization using operators with and within genetic algorithms are discussed in Azarm, Reynolds, & Narayanan [0]. The Niched Pareto genetic algorithm presented in Horn, Nafpliotis, & Goldberg [] is a MOGA that uses tournament selection of a design versus a random sample of the current population. Two individuals are selected for reproduction and each is compared to the population sample. The design that outperforms the other in the tournament is selected for reproduction. In addition to using the objective functions as a measure of fitness, a niche count is kept in each area of the performance (criterion) space. If the niche count is high in a region, a design in that region is at a disadvantage to a design in a region of low niche count. This is done to encourage a wide spread of the Pareto frontier. The metric used to quantize the niche pressure as a function of niche count can be set by the user and has a profound effect on the optimization. Some empirical evidence generated by Horn et. al. [] suggests that too much pressure results in convergence on a small portion of the entire frontier while too little pressure results in a large number of dominated solutions in the final population (recall from above that at selection time, designs are compared to only a random sample of the population, not the entire thing). Other measures of fitness are used within GA s to facilitate the generation of Pareto points. Balling [] utilizes a maxmin formulation that takes a minimum value of normalized objectives and maximizes it over all the possible designs in the current generation. This fitness function will generate nondominated designs. Other ranking schemes have also been proposed [3, 4], but ranking does not provide a way to quantitatively measure the difference in Pareto-optimality. Because of the success of genetic algorithms as effective approaches to generating Pareto frontiers, in this paper we present a MOGA catered specifically for multiobjective Copyright 00 by ASME

3 problems by taking advantage of a novel fitness function, clone testing, and introductions of epidemics. Section. presents a discussion of genetic algorithms and some details regarding the specific implementation used here.. Genetic Algorithms in Optimization Genetic Algorithms were originally developed to imitate the processes by which living beings evolve [5]. Nearly all design methodologies incorporate the concept of evolving designs. Even designers who do not subscribe to a methodology typically evolve their designs by trial and error. Genetic Algorithms are most useful for problems involving multi-modal design spaces. The fact that they do not consider any gradient information makes it possible for the algorithm to move between the peaks of a multi-modal space. A gradientbased optimizer would typically remain within a single mode throughout the solution process likely resulting in a highly suboptimal solution. A suboptimal solution is also possible when using a GA in a multi-modal space but it is less likely. It is also likely to get closer to the true optimal than the solution given by the gradient-based optimizer. One major focus of our work is on computational efficiency both in memory usage and computational time. We have tailored our Genetic Algorithm accordingly. In the following paragraphs, we describe the operators for our GA along with some of the methods we have employed to help ensure computational efficiency. Binary Encoding Each design variable is represented as a single 3 bit signed integer value within the computer. Therefore, 30 bits are available for genetic encoding. The desired decimal precision for each design variable is input by the user. Floating point numbers are converted to integers by multiplying by 0 prec(i) where prec(i) is the desired decimal precision of the i th design variable. An array of 30 short integers for genetic encoding would require at least 60 bytes or 40 bits compared to 4 bytes or 3 bits and accommodations would still be necessary for a sign and a decimal place. Another advantage of using the signed integer representation of design variables is that conversion from a binary representation to a decimal representation requires no code. It is done at the hardware level, which saves a great deal of computational time. Dynamic Memory Allocation All the arrays in the genetic algorithm are sized at either compile or run time using dynamic memory allocation commands according to the problem parameters (number of design variables, number of constraints, etc). The alternative is to allocate arbitrarily large arrays and perhaps use only a portion of them. We have incorporated the basic components of reproduction and selection in our Genetic Algorithm. They are presented and described in the following paragraphs. Generation of an Initial Population The initial population is simply a collection of designs with design variable values taken as random numbers between the design variable bounds. Each design is tested to be sure that it is unique to encourage diversity amongst the initial population. Clone Testing An efficient algorithm for duplicate design point detection is available with this GA. This is useful for 4 reasons.. To avoid design point re-evaluation.. To encourage design space exploration. 3. To avoid wasting memory to store duplicate design point information, and 4. To avoid premature convergence in a suboptimal mode. Each new design is tested to see that it is unique amongst every other design evaluated during the optimization run (including those that were discarded). Evaluation of Design Fitness A set of objective function values must be computed for each design. The form of the functions is not important to the algorithm. Either calls to other software packages or actual analytical functions are suitable. Constraints are computed for each design and used to penalize infeasible designs at selection and replacement time. A value is stored for each design that indicates how infeasible it is. The value is computed as shown in Equation (L norm). D = m j= l {max[ 0, g ( x)]} + h ( x) () j k = In equation, m is the number of inequality constraints and l is the number of equality constraints. Side constraints are handled separately. Reproduction Reproduction is accomplished in three stages. First, a mating pool is created based on the feasibility metric of Equation. All individuals are represented at least once in the mating pool. Individuals with a lower value from Equation are represented more times in the mating pool making it more likely that they will be given an opportunity to reproduce. All Feasible designs are given an equal chance of reproducing. Individuals are randomly chosen from the mating pool. The k 3 Copyright 00 by ASME

4 number of individuals selected is determined by the size of the population, the number of children produced per crossover, and the user input crossover rate. Crossover Two individuals selected for reproduction are chosen as a mating pair. Each mating pair can produce any even number of offspring (value input by user). The offspring will be referred to as children. Two types of crossover are available in this GA. The first is Single Point Uniform Parameterized Crossover [6] and is performed for each design variable as shown in Figure. Binary (encoded) Decimal Parent 03 Parent Child Child Figure : Single Point Uniform Parameterized Crossover The vertical line serves as a randomly selected crossover point. The actual bit switch is accomplished using bitwise logical operators. The second is Arithmetic Crossover and is carried out according to equations 3 and 4 below. Child = ( r)( Parent) + ( r)( Parent) (3) Child = ( r)( Parent) + ( r)( Parent) (4) In the above equations, r is U[0,]. In this way, the children are the result of a convex combination of the parents [7]. Mutation Children from the crossover stage are randomly selected for mutation. There are two types of mutation available in this GA. The first is referred to as random bit mutation. In random bit mutation, a design variable is randomly selected, as is a bit location for mutation. The selected bit is negated. If it was a 0 it becomes a and vice versa. The number of mutations is determined by the number of children created, the total number of bits making up all the children, and the user input mutation rate. The second mutation type is referred to as random design variable mutation. In this form of mutation, a design variable is chosen at random and reassigned to a random value within the upper and lower bounds for that variable. The number of mutations is determined by the number of children, the number of design variables, and the user input mutation rate. Insertion of Children into Population Once the children have been created and mutated, the process of inserting them into the population begins. This is where the Pareto dominance of the designs is considered. If there are infeasible members in the population, then the one with the largest D (from Equation ) is selected for comparison and possible replacement. If the child that it is being compared to is feasible or has a lesser D, then the child replaces it. Otherwise, the child is not allowed entrance into the population. If all members of the population are feasible and the child is feasible, then the population is searched for a member that is dominated by the child. The function that performs the search returns the design that is most dominated by the child. This is done by keeping track of the current worst design as it searches and comparing each successive design to the current worst. If it is found that the child does not dominate any members of the population, then a check is run to see if any members of the population dominate the child. If the child is dominated by a member of the population, then it is not permitted entrance into the population. If the child is not dominated by any member of the population, then the population size is increased and the child is granted entrance. Because of the way new designs are admitted into the population, a sub-optimal design has a chance of remaining within the population and reproducing for a significant number of generations. This can cause the population size to become unnecessarily large which significantly affects the speed and efficiency of the algorithm. Therefore, an additional capability is developed for this work whereby the user can periodically eliminate all dominated designs from the population. This is termed an introduction of an epidemic, as it eliminates all the unfit members of a population. The frequency at which epidemics occur is set by the user. The two parameters that define the frequency of the epidemics are the Cut-off Point and the Cut-off Point Adder. The Cut-off Point is the size that the population must reach before an epidemic is introduced. After an epidemic, the Cut-off Point is incremented by the Cut-off Point Adder and the population is allowed to expand to the new Cut-off Point before another epidemic is introduced. Pareto Set Quality Metrics Upon completion of the optimization, two quality metrics are computed for the resulting Pareto set. These quality metrics were developed by Wu & Azarm [8] because of the lack of standard metrics to evaluate the effectiveness of any method to 4 Copyright 00 by ASME

5 generate points on a Pareto frontier. The two metrics are described in the following sections. Distinct Point Metric The first metric is the number of distinct choices in the set. To compute this metric, the design space is divided into a hypergrid. A distinct choice is one that exists alone in a grid location as in Figure. third constraint limits the stress in bar BC to no more than 00,000 kpa. The three design variables are the cross sectional areas of bars AC and BC labeled x and x respectively and the overall height of the truss labeled y. Side constraints are placed on all three design variables. 4m m Pareto Distinct A B x x y Figure : Distinct Point Quality Metric The grid spacing along each objective in this work is determined by a user input Cluster Percentage and the range of objective function values in the population. Cluster Metric The second metric is the Cluster Metric. This metric gives an indication of how uniformly spaced the final population is. It is computed as the number of Pareto points divided by the number of distinct Pareto points. The target value for the cluster metric is, which corresponds to a Pareto set in which every member is distinct. The worst possible value is equal to the number of Pareto points which is the case when all the points are tightly clustered such that only a single design is considered distinct. The next section contains two case studies used to verify the approach presented above. 3 CASE STUDIES All Trials for the following case studies were carried out on a PC with a 600 MHz Pentium III processor and 56 MB of RAM. 3. Two Bar Truss This case study involves a multi-objective -bar truss problem adapted from Azarm, Reynolds, & Narayanan [0]. A diagram of the truss is shown in Figure 3. The two objectives in this problem are to minimize the overall volume of material used and to minimize the stress in bar AC (see Figure 3). The system is subject to three inequality constraints. The first two are constraints on the objective function values. They are intended to limit the size of the Pareto set. The Pareto set is unbounded if these constraints are not in place. The side constraints will limit the Pareto set but in a less concise manner than the inequality constraints. The The problem formulation is given below. Minimize: Subject to: f volume = x 6 + y f stress, AC + y + x 0 = 6 + y yx g ( x, ) f 0. y 00kN Figure 3: Truss for Case Study 3 volume g ( x, y) f stress, AC 00, y g 3( x, y) 00, 000 yx C y 3 (meters) > x > 0. (meters ) > x > 0. (meters ) 5 Copyright 00 by ASME

6 The MOGA parameters used for the optimization are given in Table. Mutation Type Random Design Variable Crossover Type Sing. Pt. Uniform Par. Clone Test Enabled Initial Population Size 00 # of Children (per crossover) Mutation Rate 7% Crossover Rate 0% Initial Cutoff Point 000 Cutoff Point Adder 500 Max # of Generations 000 Decimal Prec. (all variables) 4 Cluster % (all objectives) % Table : Optimization Parameters for -bar Truss The results of applying the multi-objective genetic algorithm presented in this paper to the truss example are presented in Table. This problem was selected because it is used by Azarm, et al [0] to compare MOGA s with repeated applications of single objective genetic algorithms. The results in Table above are compared to the results given by Azarm, et al [0]. Time of execution 5 Seconds # of non-dom. Pts # of distinct non-dom. Pts. 6 Cluster Metric # of obj. fctn. evaluations 9090 # of Generations 0 Table : Results for Optimization of -bar Truss Before presenting the comparison, it is extremely important to note that implementation in this work is not the same as that of Azarm, et. al. [0]. First, the genetic parameters used here are not the same nor are the genetic operators. For instance, Azarm uses -point uniform crossover and random bit mutation where single point uniform parameterized crossover and random design variable mutation are used in this work. Second, Azarm computes distinct solutions by computing the distance between points on the Pareto frontier and comparing the result to some limiting value. This is similar to but not the same as the grid implementation presented by Wu & Azarm [8] and also used in this work. No indication was given by Azarm regarding the limiting distance for distinct points and therefore it cannot be compared to the Cluster Percentage used here. Third, the MOGA used by Azarm incorporates niche pressure and this work does not in the approach presented in this paper. Fourth, Azarm did not supply upper bounds on the cross sectional areas ( x ). The algorithm used in this work requires them. Therefore, the upper bounds were set to very large values. Finally, the decimal precision used by Azarm was not reported and certainly has an effect on the efficiency of the optimization. With this in mind, the results are now presented for general comparison purposes in Table 3. Method Total Pts. Pareto Pts. Distinct Pts. I-SHOT 0, MOGA 3, This MOGA 9,090 3,049 6 Table 3: Comparison of Previous work to this work The I-SHOT method visited,974 points for every Pareto point found, the MOGA of Azarm visited 97 points for every Pareto point found, and the MOGA implementation of this work visited 3 points for every Pareto point found. The number of distinct points is strongly related to the limiting distance in Azarm, et. al. [0] and the Cluster percentage used in this work. Since Azarm did not report the limiting distance, the two cannot be compared. Therefore the number of distinct points should not be considered as a strong indicator of the relative efficiencies of the Azarm methods and the method presented here. Azarm also mentions that some duplication occurred and was part of the reason that there were fewer distinct Pareto points than total Pareto points. Duplication was not allowed in this work and this should be considered as a point for comparison. Based on this information, the algorithm used in this work compares favorably to the ones used by Azarm, et al [0]. There are significantly fewer evaluations per non-dominated point in both cases. It is important to note however that many of the points generated by this MOGA are non-distinct. Such a dense sampling of the Pareto set is likely unnecessary. Figure 4 contains the points generated along the Pareto frontier for this problem using the MOGA presented in this paper. 6 Copyright 00 by ASME

7 Stress in Bar AC (MPa) Pareto Frontier Volume of Material (m^3) 0. R 36.0 (inches) 0.5 T 6.0 (inches) 0. L 40.0 (inches) The values for the constants in the problem are listed in Table 4. Parameter Value P 3.89 ksi S t 35.0 ksi ρ 0.83 lbs/in 3 3. Pressure Vessel The second case study is a multi-objective pressure vessel problem taken from Winer [9]. The two objectives are to maximize the volume and to minimize the weight of a pressure vessel. The objective functions are both non-linear. There are four inequality constraints the first of which is non-linear and the rest of which are linear. The first inequality constraint limits the maximum circumferential stress in the vessel while the other three are geometric constraints. There are 3 design variables, they are the radius of the tank (R), the thickness of the tank (t), and the length of the tank (L). There are side constraints on all the design variables. All dimensions are in inches and all weights are in pounds. The problem statement is given below. Maximize: Minimize: f wgt Subject to: Figure 4: Pareto Set Generated 4 3 f vol = π R + πr L 3 = ρ π ( R + T ) + π ( R + T ) L ( πr + πr L) 3 3 PR g ( R, T) S t 0 T g ( R, ) 5T R 0 T g ( R, ) R + T T g ( R, L, ) L + R + T T Table 4: Constants for Case Study The parameters used for the optimization are given in Table 5. Mutation Type Random Design Variable Crossover Type Sing. Pt. Uniform Par. Clone Test Enabled Initial Population Size 00 # of Children (per crossover) Mutation Rate 7% Crossover Rate 0% Initial Cutoff Point 000 Cutoff Point Adder 000 Max # of Generations 000 Decimal Prec. (all variables) 5 Cluster % (all objectives) 5% Table 5: Optimization Parameters for Pressure Vessel The results of the applying the multiobjective optimization genetic algorithm presented in this paper to pressure vessel example are presented in Table 6. Time of execution 8 Seconds # of non-dom. Pts. 4 # of distinct non-dom. Pts. 39 Cluster Metric # of obj. fctn. evaluations 3,588 # of Generations 50 Table 6: Results for Optimization of Pressure Vessel The most noticeable value in Table 6 is the cluster metric. Recall from Section that the target value for the cluster metric is. This high value can be caused by one of three things. The first possible cause is that the algorithm quickly moved to distinct portions of the space and remained in these 7 Copyright 00 by ASME

8 areas creating clusters. The second possible cause is that the cluster percentage is high forcing the space between distinct points to be large and thus only a small percentage of the generated points are considered distinct. This may be desired in which case a high cluster metric is not necessarily a bad thing accepting that there were far more function evaluations than necessary. The final possible cause is that the Pareto frontier is well represented but the high decimal precision allows for many closely spaced performance points and the clustering occurs on the entire frontier (recall that no explicit niche forming was used). The algorithm was able to produce non-dominated point for every 5 points visited. Figure 5 shows all the non-dominated designs generated. The plot shows the objective space, given by Volume vs. Weight, where larger volumes are preferred and smaller weights are preferred. The data is displayed using an in-house visualization package called Cloud Visualization [0]. The visualization package allows a user to visualize various sets of points, including dominated points, non-dominated points, distinct Pareto points (see Section.) as clouds of points. Figure 6 is the same data as that of Figure 5 with all the nondistinct points filtered out. Volume (in 3 ) Desired Region Weight (lbs) Figure 5: Case Study Performance Space Nondominated Points Only Volume (in 3 ) Weight (lbs) Figure 6: Case Study Performance space distinct non-dominated points only. 4 CONCLUSIONS This problem was selected because it is used by Azarm, et al [0] to compare MOGA s with repeated applications of single objective genetic algorithms. The results in Table above are compared to the results given by Azarm, et al [0]. In this paper we have presented an efficient means of populating a Pareto frontier using genetic programming. The method used here was shown to provide a far better ratio of total points to Pareto points than other methods for the first test case presented. A thorough sampling of the Pareto frontier was also obtained for the second case study with a good ratio of total points to Pareto points. The multiple objective genetic algorithm presented takes advantage of a number of conceptual and computational developments to effectively generate Pareto frontiers quite efficiently. The use of binary encoding, dynamic memory allocation, clone testing, and epidemics allow the genetic algorithm to generate many Pareto points in one optimization process, giving a designer better information about the problem at hand. The algorithm has also been applied in parallel computational environments [] to optimize and simulate computationally expensive vehicle dynamics design problems. Future work includes the addition of niche forming to this algorithm as well as real-time visualization of the evolutionary process. ACKNOWLEDGEMENTS We acknowledge the support of the National Science Foundation, grant DMII in this work. REFERENCES [] Pareto, V., 906, Manuale di Econòmica Polìttica, Società Editrice Libràia, Milan, Italy; translated into English by A. S. Schwier, as Manual of Political Economy, Macmillan, New York, Copyright 00 by ASME

9 [] Chen. W., Wiecek, M., and Zhang, J., 999, Quality Utility: A Compromise Programming Approach to Robust Design, ASME Journal of Mechanical Design, Vol., pp [3] Das, I. and Dennis, J. E., 997, A Closer Look at Drawbacks of Minimizing Weighted Sums of Objectives for Pareto set Generation in Multicriteria Optimization Problems, Structural Optimization, Vol. 4, pp [4] Messac, A., Sundararaj, J. G., Tappeta R. V., and Renaud, J. E., 000, The Ability of Objective Functions to Generate Points on Non-Convex Pareto Frontiers, AIAA Journal, Vol. 38, pp [5] Messac, A., Melachrinoudis, E., and Sukam, C. P., 000, Aggregate Objective Functions and Pareto Frontiers: Required Relationships and Practical Implications, Optimization and Engineering, Kluwer Academic Publishers, Netherlands, Vol., pp [6] Kasprzak, E. and Lewis K., 000, An Approach to Facilitate Decision Tradeoffs in Pareto Solution Sets, Journal of Engineering Valuation and Cost Analysis, Vol. 3, pp [7] Das, I. and Dennis J. E., 998, Normal-Boundary Intersection: A New Method for Generating the Pareto Surface in Nonlinear Multicriteria Optimization Problems, SIAM Journal on Optimization, Vol. 8, pp [8] Messac, A. and Sundararaj, J. G., 000, Physical Programming s Ability to Generate a Well-Distributed Set of Pareto Points, 4 st AIAA/ASME/ASCE/AHS/ASC Structures, Structural Dynamics and Materials Conference, Paper No. AIAA [9] Narayanan, S., and Azarm, S., 999, A Multiobjective Interactive Sequential Hybrid Optimization Technique for Design Decision Making, Engineering Optimization, Vol. 3, pp [0] Azarm S., Reynolds, B. J., and Narayanan, S., 999, Comparison of Two Multiobjective Optimization Techniques with and within Genetic Algorithms, ASME Design Engineering Technical Conferences, DETC99/DAC [] Horn, J., Nafpliotis, N., and Goldberg, D. E., 994, A Niched Pareto Genetic Algorithm for Multiobjective Optimization, IEEE Conference on Evolutionary Computation, Vol., pp [] Balling, R. J., 000, Pareto Sets in Decision-Based Design, Engineering Valuation and Cost Analysis, Vol. 3, pp [3] Goldberg, D.E., 989, Genetic Algorithms in Search, Optimization, and Machine Learning, Addison-Wesley, Reading, MA. [4] Fonseca, C. M. and Fleming, P. J., 998, Multiobjective Optimization and Multiple Constraint Handling with Evolutionary Algorithms Part : A Unified Formulation, IEEE Transactions on Systems, Management, and Cybernetics Part A: Systems and Humans, Vol. 8, pp [5] Holland, J. H., 99, Adaptation in Natural and Artificial Systems, Fifth Printing, Cambridge, Mass., First MIT Press edition. [6] Anderson, K. S. and Hsu, Y., 998, Crossover Strategy For Improved Solution Space Exploration With Genetic Algorithms, ASME Design Engineering Technical Conferences, DETC98/DAC-567. [7] Gen, M. and Cheng, R., 000, Genetic Algorithms & Engineering Optimization, John Wiley & Sons, Inc., New York, NY. [8] Wu, J. and Azarm, S., 000, Metrics for Quality Assessment of a Multiobjective Design Optimization Solution Set, ASME Design Engineering Technical Conferences, DETC000/DAC-433. [9] Winer E. H., 999, Development of Visualization Techniques as an Aid in Multidisciplinary Design Optimization, Ph.D. Dissertation, Department of Mechanical and Aerospace Engineering, University at Buffalo. [0] Eddy, J., 00, The Use of Genetic Programming and Visualization to Facilitate Multi-Objective Design Optimization, M.S. Thesis, Department of Mechanical and Aerospace Engineering, University at Buffalo. [] Eddy, J., K. Hacker, and K. Lewis, 000, Solving Computationally Expensive Optimization Problems Using Hybrid Methods in Parallel Computing Environments, 8th AIAA/USAF/NASA/ISSMO Symposium on Multidisciplinary Analysis and Optimization, Long Beach, CA, AIAA Copyright 00 by ASME

Multi-objective Optimization

Multi-objective Optimization Jugal K. Kalita Single vs. Single vs. Single Objective Optimization: When an optimization problem involves only one objective function, the task of finding the optimal solution is called single-objective

More information

Multi-objective Optimization

Multi-objective Optimization Some introductory figures from : Deb Kalyanmoy, Multi-Objective Optimization using Evolutionary Algorithms, Wiley 2001 Multi-objective Optimization Implementation of Constrained GA Based on NSGA-II Optimization

More information

CHAPTER 6 REAL-VALUED GENETIC ALGORITHMS

CHAPTER 6 REAL-VALUED GENETIC ALGORITHMS CHAPTER 6 REAL-VALUED GENETIC ALGORITHMS 6.1 Introduction Gradient-based algorithms have some weaknesses relative to engineering optimization. Specifically, it is difficult to use gradient-based algorithms

More information

Exploring Mass Trade-Offs In Preliminary Vehicle Design Using Pareto Sets

Exploring Mass Trade-Offs In Preliminary Vehicle Design Using Pareto Sets Exploring Mass Trade-Offs In Preliminary Vehicle Design Using Pareto Sets Joseph Donndelinger 1 General Motors Research & Development Center, Warren, Michigan 48090 Scott Ferguson 2 and Kemper Lewis 3

More information

Suppose you have a problem You don t know how to solve it What can you do? Can you use a computer to somehow find a solution for you?

Suppose you have a problem You don t know how to solve it What can you do? Can you use a computer to somehow find a solution for you? Gurjit Randhawa Suppose you have a problem You don t know how to solve it What can you do? Can you use a computer to somehow find a solution for you? This would be nice! Can it be done? A blind generate

More information

Exploration of Pareto Frontier Using a Fuzzy Controlled Hybrid Line Search

Exploration of Pareto Frontier Using a Fuzzy Controlled Hybrid Line Search Seventh International Conference on Hybrid Intelligent Systems Exploration of Pareto Frontier Using a Fuzzy Controlled Hybrid Line Search Crina Grosan and Ajith Abraham Faculty of Information Technology,

More information

Generating Uniformly Distributed Pareto Optimal Points for Constrained and Unconstrained Multicriteria Optimization

Generating Uniformly Distributed Pareto Optimal Points for Constrained and Unconstrained Multicriteria Optimization Generating Uniformly Distributed Pareto Optimal Points for Constrained and Unconstrained Multicriteria Optimization Crina Grosan Department of Computer Science Babes-Bolyai University Cluj-Napoca, Romania

More information

Revision of a Floating-Point Genetic Algorithm GENOCOP V for Nonlinear Programming Problems

Revision of a Floating-Point Genetic Algorithm GENOCOP V for Nonlinear Programming Problems 4 The Open Cybernetics and Systemics Journal, 008,, 4-9 Revision of a Floating-Point Genetic Algorithm GENOCOP V for Nonlinear Programming Problems K. Kato *, M. Sakawa and H. Katagiri Department of Artificial

More information

An Evolutionary Algorithm for the Multi-objective Shortest Path Problem

An Evolutionary Algorithm for the Multi-objective Shortest Path Problem An Evolutionary Algorithm for the Multi-objective Shortest Path Problem Fangguo He Huan Qi Qiong Fan Institute of Systems Engineering, Huazhong University of Science & Technology, Wuhan 430074, P. R. China

More information

Constrained Multi-Objective Optimization of a Condenser Coil Using Evolutionary Algorithms

Constrained Multi-Objective Optimization of a Condenser Coil Using Evolutionary Algorithms Purdue University Purdue e-pubs International Refrigeration and Air Conditioning Conference School of Mechanical Engineering 2004 Constrained Multi-Objective Optimization of a Condenser Coil Using Evolutionary

More information

Mechanical Component Design for Multiple Objectives Using Elitist Non-Dominated Sorting GA

Mechanical Component Design for Multiple Objectives Using Elitist Non-Dominated Sorting GA Mechanical Component Design for Multiple Objectives Using Elitist Non-Dominated Sorting GA Kalyanmoy Deb, Amrit Pratap, and Subrajyoti Moitra Kanpur Genetic Algorithms Laboratory (KanGAL) Indian Institute

More information

Heuristic Optimisation

Heuristic Optimisation Heuristic Optimisation Part 10: Genetic Algorithm Basics Sándor Zoltán Németh http://web.mat.bham.ac.uk/s.z.nemeth s.nemeth@bham.ac.uk University of Birmingham S Z Németh (s.nemeth@bham.ac.uk) Heuristic

More information

Multiobjective Optimization (II)

Multiobjective Optimization (II) Multidisciplinary System Design Optimization (MSDO) Multiobjective Optimization (II) Lecture 15 Dr. Anas Alfaris 1 Massachusetts Institute of Technology - Prof. de Weck and Prof. Willcox MOO 2 Lecture

More information

MINIMAL EDGE-ORDERED SPANNING TREES USING A SELF-ADAPTING GENETIC ALGORITHM WITH MULTIPLE GENOMIC REPRESENTATIONS

MINIMAL EDGE-ORDERED SPANNING TREES USING A SELF-ADAPTING GENETIC ALGORITHM WITH MULTIPLE GENOMIC REPRESENTATIONS Proceedings of Student/Faculty Research Day, CSIS, Pace University, May 5 th, 2006 MINIMAL EDGE-ORDERED SPANNING TREES USING A SELF-ADAPTING GENETIC ALGORITHM WITH MULTIPLE GENOMIC REPRESENTATIONS Richard

More information

Assessing the Convergence Properties of NSGA-II for Direct Crashworthiness Optimization

Assessing the Convergence Properties of NSGA-II for Direct Crashworthiness Optimization 10 th International LS-DYNA Users Conference Opitmization (1) Assessing the Convergence Properties of NSGA-II for Direct Crashworthiness Optimization Guangye Li 1, Tushar Goel 2, Nielen Stander 2 1 IBM

More information

Evolutionary Algorithms: Lecture 4. Department of Cybernetics, CTU Prague.

Evolutionary Algorithms: Lecture 4. Department of Cybernetics, CTU Prague. Evolutionary Algorithms: Lecture 4 Jiří Kubaĺık Department of Cybernetics, CTU Prague http://labe.felk.cvut.cz/~posik/xe33scp/ pmulti-objective Optimization :: Many real-world problems involve multiple

More information

Recombination of Similar Parents in EMO Algorithms

Recombination of Similar Parents in EMO Algorithms H. Ishibuchi and K. Narukawa, Recombination of parents in EMO algorithms, Lecture Notes in Computer Science 341: Evolutionary Multi-Criterion Optimization, pp. 265-279, Springer, Berlin, March 25. (Proc.

More information

Genetic Algorithm Performance with Different Selection Methods in Solving Multi-Objective Network Design Problem

Genetic Algorithm Performance with Different Selection Methods in Solving Multi-Objective Network Design Problem etic Algorithm Performance with Different Selection Methods in Solving Multi-Objective Network Design Problem R. O. Oladele Department of Computer Science University of Ilorin P.M.B. 1515, Ilorin, NIGERIA

More information

The Genetic Algorithm for finding the maxima of single-variable functions

The Genetic Algorithm for finding the maxima of single-variable functions Research Inventy: International Journal Of Engineering And Science Vol.4, Issue 3(March 2014), PP 46-54 Issn (e): 2278-4721, Issn (p):2319-6483, www.researchinventy.com The Genetic Algorithm for finding

More information

IMPROVING A GREEDY DNA MOTIF SEARCH USING A MULTIPLE GENOMIC SELF-ADAPTATING GENETIC ALGORITHM

IMPROVING A GREEDY DNA MOTIF SEARCH USING A MULTIPLE GENOMIC SELF-ADAPTATING GENETIC ALGORITHM Proceedings of Student/Faculty Research Day, CSIS, Pace University, May 4th, 2007 IMPROVING A GREEDY DNA MOTIF SEARCH USING A MULTIPLE GENOMIC SELF-ADAPTATING GENETIC ALGORITHM Michael L. Gargano, mgargano@pace.edu

More information

Genetic algorithms and finite element coupling for mechanical optimization

Genetic algorithms and finite element coupling for mechanical optimization Computer Aided Optimum Design in Engineering X 87 Genetic algorithms and finite element coupling for mechanical optimization G. Corriveau, R. Guilbault & A. Tahan Department of Mechanical Engineering,

More information

HYBRID GENETIC ALGORITHM WITH GREAT DELUGE TO SOLVE CONSTRAINED OPTIMIZATION PROBLEMS

HYBRID GENETIC ALGORITHM WITH GREAT DELUGE TO SOLVE CONSTRAINED OPTIMIZATION PROBLEMS HYBRID GENETIC ALGORITHM WITH GREAT DELUGE TO SOLVE CONSTRAINED OPTIMIZATION PROBLEMS NABEEL AL-MILLI Financial and Business Administration and Computer Science Department Zarqa University College Al-Balqa'

More information

Metaheuristic Optimization with Evolver, Genocop and OptQuest

Metaheuristic Optimization with Evolver, Genocop and OptQuest Metaheuristic Optimization with Evolver, Genocop and OptQuest MANUEL LAGUNA Graduate School of Business Administration University of Colorado, Boulder, CO 80309-0419 Manuel.Laguna@Colorado.EDU Last revision:

More information

ANTICIPATORY VERSUS TRADITIONAL GENETIC ALGORITHM

ANTICIPATORY VERSUS TRADITIONAL GENETIC ALGORITHM Anticipatory Versus Traditional Genetic Algorithm ANTICIPATORY VERSUS TRADITIONAL GENETIC ALGORITHM ABSTRACT Irina Mocanu 1 Eugenia Kalisz 2 This paper evaluates the performances of a new type of genetic

More information

Lamarckian Repair and Darwinian Repair in EMO Algorithms for Multiobjective 0/1 Knapsack Problems

Lamarckian Repair and Darwinian Repair in EMO Algorithms for Multiobjective 0/1 Knapsack Problems Repair and Repair in EMO Algorithms for Multiobjective 0/ Knapsack Problems Shiori Kaige, Kaname Narukawa, and Hisao Ishibuchi Department of Industrial Engineering, Osaka Prefecture University, - Gakuen-cho,

More information

Meta- Heuristic based Optimization Algorithms: A Comparative Study of Genetic Algorithm and Particle Swarm Optimization

Meta- Heuristic based Optimization Algorithms: A Comparative Study of Genetic Algorithm and Particle Swarm Optimization 2017 2 nd International Electrical Engineering Conference (IEEC 2017) May. 19 th -20 th, 2017 at IEP Centre, Karachi, Pakistan Meta- Heuristic based Optimization Algorithms: A Comparative Study of Genetic

More information

Comparison of Evolutionary Multiobjective Optimization with Reference Solution-Based Single-Objective Approach

Comparison of Evolutionary Multiobjective Optimization with Reference Solution-Based Single-Objective Approach Comparison of Evolutionary Multiobjective Optimization with Reference Solution-Based Single-Objective Approach Hisao Ishibuchi Graduate School of Engineering Osaka Prefecture University Sakai, Osaka 599-853,

More information

CHAPTER 2 CONVENTIONAL AND NON-CONVENTIONAL TECHNIQUES TO SOLVE ORPD PROBLEM

CHAPTER 2 CONVENTIONAL AND NON-CONVENTIONAL TECHNIQUES TO SOLVE ORPD PROBLEM 20 CHAPTER 2 CONVENTIONAL AND NON-CONVENTIONAL TECHNIQUES TO SOLVE ORPD PROBLEM 2.1 CLASSIFICATION OF CONVENTIONAL TECHNIQUES Classical optimization methods can be classified into two distinct groups:

More information

An Approach to Facilitate Decision Tradeoffs in Pareto Solution Sets

An Approach to Facilitate Decision Tradeoffs in Pareto Solution Sets Published in Journal of Engineering Valuation and Cost Analysis, Vol. 3, No. 1, pp. 173-187, 2000. An Approach to Facilitate Decision Tradeoffs in Pareto Solution Sets Edward M. Kasprzak 1 Department of

More information

Incorporation of Scalarizing Fitness Functions into Evolutionary Multiobjective Optimization Algorithms

Incorporation of Scalarizing Fitness Functions into Evolutionary Multiobjective Optimization Algorithms H. Ishibuchi, T. Doi, and Y. Nojima, Incorporation of scalarizing fitness functions into evolutionary multiobjective optimization algorithms, Lecture Notes in Computer Science 4193: Parallel Problem Solving

More information

Metaheuristic Development Methodology. Fall 2009 Instructor: Dr. Masoud Yaghini

Metaheuristic Development Methodology. Fall 2009 Instructor: Dr. Masoud Yaghini Metaheuristic Development Methodology Fall 2009 Instructor: Dr. Masoud Yaghini Phases and Steps Phases and Steps Phase 1: Understanding Problem Step 1: State the Problem Step 2: Review of Existing Solution

More information

A Similarity-Based Mating Scheme for Evolutionary Multiobjective Optimization

A Similarity-Based Mating Scheme for Evolutionary Multiobjective Optimization A Similarity-Based Mating Scheme for Evolutionary Multiobjective Optimization Hisao Ishibuchi and Youhei Shibata Department of Industrial Engineering, Osaka Prefecture University, - Gakuen-cho, Sakai,

More information

DETERMINING MAXIMUM/MINIMUM VALUES FOR TWO- DIMENTIONAL MATHMATICLE FUNCTIONS USING RANDOM CREOSSOVER TECHNIQUES

DETERMINING MAXIMUM/MINIMUM VALUES FOR TWO- DIMENTIONAL MATHMATICLE FUNCTIONS USING RANDOM CREOSSOVER TECHNIQUES DETERMINING MAXIMUM/MINIMUM VALUES FOR TWO- DIMENTIONAL MATHMATICLE FUNCTIONS USING RANDOM CREOSSOVER TECHNIQUES SHIHADEH ALQRAINY. Department of Software Engineering, Albalqa Applied University. E-mail:

More information

Grid Scheduling Strategy using GA (GSSGA)

Grid Scheduling Strategy using GA (GSSGA) F Kurus Malai Selvi et al,int.j.computer Technology & Applications,Vol 3 (5), 8-86 ISSN:2229-693 Grid Scheduling Strategy using GA () Dr.D.I.George Amalarethinam Director-MCA & Associate Professor of Computer

More information

INTERACTIVE MULTI-OBJECTIVE GENETIC ALGORITHMS FOR THE BUS DRIVER SCHEDULING PROBLEM

INTERACTIVE MULTI-OBJECTIVE GENETIC ALGORITHMS FOR THE BUS DRIVER SCHEDULING PROBLEM Advanced OR and AI Methods in Transportation INTERACTIVE MULTI-OBJECTIVE GENETIC ALGORITHMS FOR THE BUS DRIVER SCHEDULING PROBLEM Jorge PINHO DE SOUSA 1, Teresa GALVÃO DIAS 1, João FALCÃO E CUNHA 1 Abstract.

More information

ISSN: [Keswani* et al., 7(1): January, 2018] Impact Factor: 4.116

ISSN: [Keswani* et al., 7(1): January, 2018] Impact Factor: 4.116 IJESRT INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH TECHNOLOGY AUTOMATIC TEST CASE GENERATION FOR PERFORMANCE ENHANCEMENT OF SOFTWARE THROUGH GENETIC ALGORITHM AND RANDOM TESTING Bright Keswani,

More information

Using Database Storage to Improve Explorative Optimization of Form Critical Structures

Using Database Storage to Improve Explorative Optimization of Form Critical Structures 15 to 19 September 2014, Brasilia, Brazil Reyolando M.L.R.F. BRASIL and Ruy M.O. PAULETTI (eds.) Using Database Storage to Improve Explorative Optimization of Form Critical Structures VON BUELOW, Peter

More information

Genetic algorithm based on number of children and height task for multiprocessor task Scheduling

Genetic algorithm based on number of children and height task for multiprocessor task Scheduling Genetic algorithm based on number of children and height task for multiprocessor task Scheduling Marjan Abdeyazdan 1,Vahid Arjmand 2,Amir masoud Rahmani 3, Hamid Raeis ghanavati 4 1 Department of Computer

More information

Genetic Algorithms Variations and Implementation Issues

Genetic Algorithms Variations and Implementation Issues Genetic Algorithms Variations and Implementation Issues CS 431 Advanced Topics in AI Classic Genetic Algorithms GAs as proposed by Holland had the following properties: Randomly generated population Binary

More information

Minimum Weight Optimization of a Gear Train by using Genetic Algorithm

Minimum Weight Optimization of a Gear Train by using Genetic Algorithm International Journal of Current Engineering and Technology E-ISSN 2277 416, P-ISSN 2347 5161 216 INPRESSCO, All Rights Reserved Available at http://inpressco.com/category/ijcet Research Article Minimum

More information

Bi-Objective Optimization for Scheduling in Heterogeneous Computing Systems

Bi-Objective Optimization for Scheduling in Heterogeneous Computing Systems Bi-Objective Optimization for Scheduling in Heterogeneous Computing Systems Tony Maciejewski, Kyle Tarplee, Ryan Friese, and Howard Jay Siegel Department of Electrical and Computer Engineering Colorado

More information

DERIVATIVE-FREE OPTIMIZATION

DERIVATIVE-FREE OPTIMIZATION DERIVATIVE-FREE OPTIMIZATION Main bibliography J.-S. Jang, C.-T. Sun and E. Mizutani. Neuro-Fuzzy and Soft Computing: A Computational Approach to Learning and Machine Intelligence. Prentice Hall, New Jersey,

More information

Topological Machining Fixture Layout Synthesis Using Genetic Algorithms

Topological Machining Fixture Layout Synthesis Using Genetic Algorithms Topological Machining Fixture Layout Synthesis Using Genetic Algorithms Necmettin Kaya Uludag University, Mechanical Eng. Department, Bursa, Turkey Ferruh Öztürk Uludag University, Mechanical Eng. Department,

More information

The Normalized Normal Constraint Method for Generating the Pareto Frontier

The Normalized Normal Constraint Method for Generating the Pareto Frontier The Normalized Normal Constraint Method for Generating the Pareto Frontier Achille Messac Amir Ismail-Yahaya Christopher A. Mattson Corresponding Author Prof. Achille Messac Mechanical, Aerospace, and

More information

Mechanical Component Design for Multiple Objectives Using Elitist Non-Dominated Sorting GA

Mechanical Component Design for Multiple Objectives Using Elitist Non-Dominated Sorting GA Mechanical Component Design for Multiple Objectives Using Elitist Non-Dominated Sorting GA Kalyanmoy Deb, Amrit Pratap, and Subrajyoti Moitra Kanpur Genetic Algorithms Laboratory (KanGAL) Indian Institute

More information

An Introduction to Evolutionary Algorithms

An Introduction to Evolutionary Algorithms An Introduction to Evolutionary Algorithms Karthik Sindhya, PhD Postdoctoral Researcher Industrial Optimization Group Department of Mathematical Information Technology Karthik.sindhya@jyu.fi http://users.jyu.fi/~kasindhy/

More information

AIRFOIL SHAPE OPTIMIZATION USING EVOLUTIONARY ALGORITHMS

AIRFOIL SHAPE OPTIMIZATION USING EVOLUTIONARY ALGORITHMS AIRFOIL SHAPE OPTIMIZATION USING EVOLUTIONARY ALGORITHMS Emre Alpman Graduate Research Assistant Aerospace Engineering Department Pennstate University University Park, PA, 6802 Abstract A new methodology

More information

OPTIMIZATION METHODS. For more information visit: or send an to:

OPTIMIZATION METHODS. For more information visit:  or send an  to: OPTIMIZATION METHODS modefrontier is a registered product of ESTECO srl Copyright ESTECO srl 1999-2007 For more information visit: www.esteco.com or send an e-mail to: modefrontier@esteco.com NEOS Optimization

More information

NODE PLACEMENT FOR A WIRELESS SENSOR NETWORK USING A MULTIOBJECTIVE GENETIC ALGORITHM

NODE PLACEMENT FOR A WIRELESS SENSOR NETWORK USING A MULTIOBJECTIVE GENETIC ALGORITHM NODE PLACEMENT FOR A WIRELESS SENSOR NETWORK USING A MULTIOBJECTIVE GENETIC ALGORITHM DAMIEN JOURDAN, PhD candidate, Department of Aeronautics and Astronautics Massachusetts Institute of Technology Cambridge,

More information

A Genetic Algorithm for Graph Matching using Graph Node Characteristics 1 2

A Genetic Algorithm for Graph Matching using Graph Node Characteristics 1 2 Chapter 5 A Genetic Algorithm for Graph Matching using Graph Node Characteristics 1 2 Graph Matching has attracted the exploration of applying new computing paradigms because of the large number of applications

More information

An Evolutionary Algorithm with Stochastic Hill-Climbing for the Edge-Biconnectivity Augmentation Problem

An Evolutionary Algorithm with Stochastic Hill-Climbing for the Edge-Biconnectivity Augmentation Problem An Evolutionary Algorithm with Stochastic Hill-Climbing for the Edge-Biconnectivity Augmentation Problem Ivana Ljubić and Günther R. Raidl Institute for Computer Graphics and Algorithms, Vienna University

More information

Mutation in Compressed Encoding in Estimation of Distribution Algorithm

Mutation in Compressed Encoding in Estimation of Distribution Algorithm Mutation in Compressed Encoding in Estimation of Distribution Algorithm Orawan Watchanupaporn, Worasait Suwannik Department of Computer Science asetsart University Bangkok, Thailand orawan.liu@gmail.com,

More information

CHAPTER 2 MULTI-OBJECTIVE REACTIVE POWER OPTIMIZATION

CHAPTER 2 MULTI-OBJECTIVE REACTIVE POWER OPTIMIZATION 19 CHAPTER 2 MULTI-OBJECTIE REACTIE POWER OPTIMIZATION 2.1 INTRODUCTION In this chapter, a fundamental knowledge of the Multi-Objective Optimization (MOO) problem and the methods to solve are presented.

More information

NCGA : Neighborhood Cultivation Genetic Algorithm for Multi-Objective Optimization Problems

NCGA : Neighborhood Cultivation Genetic Algorithm for Multi-Objective Optimization Problems : Neighborhood Cultivation Genetic Algorithm for Multi-Objective Optimization Problems Shinya Watanabe Graduate School of Engineering, Doshisha University 1-3 Tatara Miyakodani,Kyo-tanabe, Kyoto, 10-031,

More information

Extending MATLAB and GA to Solve Job Shop Manufacturing Scheduling Problems

Extending MATLAB and GA to Solve Job Shop Manufacturing Scheduling Problems Extending MATLAB and GA to Solve Job Shop Manufacturing Scheduling Problems Hamidullah Khan Niazi 1, Sun Hou-Fang 2, Zhang Fa-Ping 3, Riaz Ahmed 4 ( 1, 4 National University of Sciences and Technology

More information

A GENETIC ALGORITHM APPROACH TO OPTIMAL TOPOLOGICAL DESIGN OF ALL TERMINAL NETWORKS

A GENETIC ALGORITHM APPROACH TO OPTIMAL TOPOLOGICAL DESIGN OF ALL TERMINAL NETWORKS A GENETIC ALGORITHM APPROACH TO OPTIMAL TOPOLOGICAL DESIGN OF ALL TERMINAL NETWORKS BERNA DENGIZ AND FULYA ALTIPARMAK Department of Industrial Engineering Gazi University, Ankara, TURKEY 06570 ALICE E.

More information

Evolutionary multi-objective algorithm design issues

Evolutionary multi-objective algorithm design issues Evolutionary multi-objective algorithm design issues Karthik Sindhya, PhD Postdoctoral Researcher Industrial Optimization Group Department of Mathematical Information Technology Karthik.sindhya@jyu.fi

More information

Dynamic Uniform Scaling for Multiobjective Genetic Algorithms

Dynamic Uniform Scaling for Multiobjective Genetic Algorithms Dynamic Uniform Scaling for Multiobjective Genetic Algorithms Gerulf K. M. Pedersen 1 and David E. Goldberg 2 1 Aalborg University, Department of Control Engineering, Fredrik Bajers Vej 7, DK-922 Aalborg

More information

GENETIC ALGORITHM with Hands-On exercise

GENETIC ALGORITHM with Hands-On exercise GENETIC ALGORITHM with Hands-On exercise Adopted From Lecture by Michael Negnevitsky, Electrical Engineering & Computer Science University of Tasmania 1 Objective To understand the processes ie. GAs Basic

More information

Ajay Sharma Gaurav Kapur VK Kaushik LC Mangal RC Agarwal. Defence Electronics Applications Laboratory, Dehradun DRDO, India

Ajay Sharma Gaurav Kapur VK Kaushik LC Mangal RC Agarwal. Defence Electronics Applications Laboratory, Dehradun DRDO, India Ajay Sharma Gaurav Kapur VK Kaushik LC Mangal RC Agarwal Defence Electronics Applications Laboratory, Dehradun DRDO, India Problem considered Given a SDR with a set of configurable parameters, user specified

More information

Evolving Efficient Security Systems Under Budget Constraints Using Genetic Algorithms

Evolving Efficient Security Systems Under Budget Constraints Using Genetic Algorithms Proceedings of Student Research Day, CSIS, Pace University, May 9th, 2003 Evolving Efficient Security Systems Under Budget Constraints Using Genetic Algorithms Michael L. Gargano, William Edelson, Paul

More information

Multi-Objective Optimization using Evolutionary Algorithms

Multi-Objective Optimization using Evolutionary Algorithms Multi-Objective Optimization using Evolutionary Algorithms Kalyanmoy Deb Department of Mechanical Engineering, Indian Institute of Technology, Kanpur, India JOHN WILEY & SONS, LTD Chichester New York Weinheim

More information

A New Selection Operator - CSM in Genetic Algorithms for Solving the TSP

A New Selection Operator - CSM in Genetic Algorithms for Solving the TSP A New Selection Operator - CSM in Genetic Algorithms for Solving the TSP Wael Raef Alkhayri Fahed Al duwairi High School Aljabereyah, Kuwait Suhail Sami Owais Applied Science Private University Amman,

More information

ME 575: Two-bar Truss

ME 575: Two-bar Truss Name ME 575: Two-bar Truss Consider the design of a simple tubular symmetric truss shown in Fig. 1.1 below (problem originally from Fox 1 ). A design of the truss is specified by a unique set of values

More information

Using Genetic Algorithms in Integer Programming for Decision Support

Using Genetic Algorithms in Integer Programming for Decision Support Doi:10.5901/ajis.2014.v3n6p11 Abstract Using Genetic Algorithms in Integer Programming for Decision Support Dr. Youcef Souar Omar Mouffok Taher Moulay University Saida, Algeria Email:Syoucef12@yahoo.fr

More information

Evolutionary Computation Algorithms for Cryptanalysis: A Study

Evolutionary Computation Algorithms for Cryptanalysis: A Study Evolutionary Computation Algorithms for Cryptanalysis: A Study Poonam Garg Information Technology and Management Dept. Institute of Management Technology Ghaziabad, India pgarg@imt.edu Abstract The cryptanalysis

More information

Using Genetic Algorithms to Solve the Box Stacking Problem

Using Genetic Algorithms to Solve the Box Stacking Problem Using Genetic Algorithms to Solve the Box Stacking Problem Jenniffer Estrada, Kris Lee, Ryan Edgar October 7th, 2010 Abstract The box stacking or strip stacking problem is exceedingly difficult to solve

More information

Research Article Path Planning Using a Hybrid Evolutionary Algorithm Based on Tree Structure Encoding

Research Article Path Planning Using a Hybrid Evolutionary Algorithm Based on Tree Structure Encoding e Scientific World Journal, Article ID 746260, 8 pages http://dx.doi.org/10.1155/2014/746260 Research Article Path Planning Using a Hybrid Evolutionary Algorithm Based on Tree Structure Encoding Ming-Yi

More information

A Search Method with User s Preference Direction using Reference Lines

A Search Method with User s Preference Direction using Reference Lines A Search Method with User s Preference Direction using Reference Lines Tomohiro Yoshikawa Graduate School of Engineering, Nagoya University, Nagoya, Japan, {yoshikawa}@cse.nagoya-u.ac.jp Abstract Recently,

More information

Multi-Objective Optimization using Evolutionary Algorithms

Multi-Objective Optimization using Evolutionary Algorithms Multi-Objective Optimization using Evolutionary Algorithms Kalyanmoy Deb Department ofmechanical Engineering, Indian Institute of Technology, Kanpur, India JOHN WILEY & SONS, LTD Chichester New York Weinheim

More information

Neural Network Weight Selection Using Genetic Algorithms

Neural Network Weight Selection Using Genetic Algorithms Neural Network Weight Selection Using Genetic Algorithms David Montana presented by: Carl Fink, Hongyi Chen, Jack Cheng, Xinglong Li, Bruce Lin, Chongjie Zhang April 12, 2005 1 Neural Networks Neural networks

More information

A GENETIC ALGORITHM APPROACH FOR TECHNOLOGY CHARACTERIZATION. A Thesis EDGAR GALVAN

A GENETIC ALGORITHM APPROACH FOR TECHNOLOGY CHARACTERIZATION. A Thesis EDGAR GALVAN A GENETIC ALGORITHM APPROACH FOR TECHNOLOGY CHARACTERIZATION A Thesis by EDGAR GALVAN Submitted to the Office of Graduate Studies of Texas A&M University in partial fulfillment of the requirements for

More information

The Binary Genetic Algorithm. Universidad de los Andes-CODENSA

The Binary Genetic Algorithm. Universidad de los Andes-CODENSA The Binary Genetic Algorithm Universidad de los Andes-CODENSA 1. Genetic Algorithms: Natural Selection on a Computer Figure 1 shows the analogy between biological i l evolution and a binary GA. Both start

More information

Genetic Algorithm for Circuit Partitioning

Genetic Algorithm for Circuit Partitioning Genetic Algorithm for Circuit Partitioning ZOLTAN BARUCH, OCTAVIAN CREŢ, KALMAN PUSZTAI Computer Science Department, Technical University of Cluj-Napoca, 26, Bariţiu St., 3400 Cluj-Napoca, Romania {Zoltan.Baruch,

More information

Evolutionary Algorithms. CS Evolutionary Algorithms 1

Evolutionary Algorithms. CS Evolutionary Algorithms 1 Evolutionary Algorithms CS 478 - Evolutionary Algorithms 1 Evolutionary Computation/Algorithms Genetic Algorithms l Simulate natural evolution of structures via selection and reproduction, based on performance

More information

The Simple Genetic Algorithm Performance: A Comparative Study on the Operators Combination

The Simple Genetic Algorithm Performance: A Comparative Study on the Operators Combination INFOCOMP 20 : The First International Conference on Advanced Communications and Computation The Simple Genetic Algorithm Performance: A Comparative Study on the Operators Combination Delmar Broglio Carvalho,

More information

A New Efficient and Useful Robust Optimization Approach Design for Multi-Objective Six Sigma

A New Efficient and Useful Robust Optimization Approach Design for Multi-Objective Six Sigma A New Efficient and Useful Robust Optimization Approach Design for Multi-Objective Six Sigma Koji Shimoyama Department of Aeronautics and Astronautics University of Tokyo 3-1-1 Yoshinodai Sagamihara, Kanagawa,

More information

Evolutionary Multi-objective Optimization of Business Process Designs with Pre-processing

Evolutionary Multi-objective Optimization of Business Process Designs with Pre-processing Evolutionary Multi-objective Optimization of Business Process Designs with Pre-processing Kostas Georgoulakos Department of Applied Informatics University of Macedonia Thessaloniki, Greece mai16027@uom.edu.gr

More information

Network Routing Protocol using Genetic Algorithms

Network Routing Protocol using Genetic Algorithms International Journal of Electrical & Computer Sciences IJECS-IJENS Vol:0 No:02 40 Network Routing Protocol using Genetic Algorithms Gihan Nagib and Wahied G. Ali Abstract This paper aims to develop a

More information

REAL-CODED GENETIC ALGORITHMS CONSTRAINED OPTIMIZATION. Nedim TUTKUN

REAL-CODED GENETIC ALGORITHMS CONSTRAINED OPTIMIZATION. Nedim TUTKUN REAL-CODED GENETIC ALGORITHMS CONSTRAINED OPTIMIZATION Nedim TUTKUN nedimtutkun@gmail.com Outlines Unconstrained Optimization Ackley s Function GA Approach for Ackley s Function Nonlinear Programming Penalty

More information

CHAPTER 5 STRUCTURAL OPTIMIZATION OF SWITCHED RELUCTANCE MACHINE

CHAPTER 5 STRUCTURAL OPTIMIZATION OF SWITCHED RELUCTANCE MACHINE 89 CHAPTER 5 STRUCTURAL OPTIMIZATION OF SWITCHED RELUCTANCE MACHINE 5.1 INTRODUCTION Nowadays a great attention has been devoted in the literature towards the main components of electric and hybrid electric

More information

Multicriterial Optimization Using Genetic Algorithm

Multicriterial Optimization Using Genetic Algorithm Multicriterial Optimization Using Genetic Algorithm 180 175 170 165 Fitness 160 155 150 145 140 Best Fitness Mean Fitness 135 130 0 Page 1 100 200 300 Generations 400 500 600 Contents Optimization, Local

More information

The Parallel Software Design Process. Parallel Software Design

The Parallel Software Design Process. Parallel Software Design Parallel Software Design The Parallel Software Design Process Deborah Stacey, Chair Dept. of Comp. & Info Sci., University of Guelph dastacey@uoguelph.ca Why Parallel? Why NOT Parallel? Why Talk about

More information

Time Complexity Analysis of the Genetic Algorithm Clustering Method

Time Complexity Analysis of the Genetic Algorithm Clustering Method Time Complexity Analysis of the Genetic Algorithm Clustering Method Z. M. NOPIAH, M. I. KHAIRIR, S. ABDULLAH, M. N. BAHARIN, and A. ARIFIN Department of Mechanical and Materials Engineering Universiti

More information

Automata Construct with Genetic Algorithm

Automata Construct with Genetic Algorithm Automata Construct with Genetic Algorithm Vít Fábera Department of Informatics and Telecommunication, Faculty of Transportation Sciences, Czech Technical University, Konviktská 2, Praha, Czech Republic,

More information

GENERATIONAL MODEL GENETIC ALGORITHM FOR REAL WORLD SET PARTITIONING PROBLEMS

GENERATIONAL MODEL GENETIC ALGORITHM FOR REAL WORLD SET PARTITIONING PROBLEMS International Journal of Electronic Commerce Studies Vol.4, No.1, pp. 33-46, 2013 doi: 10.7903/ijecs.1138 GENERATIONAL MODEL GENETIC ALGORITHM FOR REAL WORLD SET PARTITIONING PROBLEMS Chi-san Althon Lin

More information

Solving A Nonlinear Side Constrained Transportation Problem. by Using Spanning Tree-based Genetic Algorithm. with Fuzzy Logic Controller

Solving A Nonlinear Side Constrained Transportation Problem. by Using Spanning Tree-based Genetic Algorithm. with Fuzzy Logic Controller Solving A Nonlinear Side Constrained Transportation Problem by Using Spanning Tree-based Genetic Algorithm with Fuzzy Logic Controller Yasuhiro Tsujimura *, Mitsuo Gen ** and Admi Syarif **,*** * Department

More information

Search direction improvement for gradient-based optimization problems

Search direction improvement for gradient-based optimization problems Computer Aided Optimum Design in Engineering IX 3 Search direction improvement for gradient-based optimization problems S Ganguly & W L Neu Aerospace and Ocean Engineering, Virginia Tech, USA Abstract

More information

What is GOSET? GOSET stands for Genetic Optimization System Engineering Tool

What is GOSET? GOSET stands for Genetic Optimization System Engineering Tool Lecture 5: GOSET 1 What is GOSET? GOSET stands for Genetic Optimization System Engineering Tool GOSET is a MATLAB based genetic algorithm toolbox for solving optimization problems 2 GOSET Features Wide

More information

Calc Redirection : A Structure for Direction Finding Aided Traffic Monitoring

Calc Redirection : A Structure for Direction Finding Aided Traffic Monitoring Calc Redirection : A Structure for Direction Finding Aided Traffic Monitoring Paparao Sanapathi MVGR College of engineering vizianagaram, AP P. Satheesh, M. Tech,Ph. D MVGR College of engineering vizianagaram,

More information

Finding Tradeoffs by Using Multiobjective Optimization Algorithms

Finding Tradeoffs by Using Multiobjective Optimization Algorithms Finding Tradeoffs by Using Multiobjective Optimization Algorithms Shigeru Obayashi, Daisuke Sasaki,* Akira Oyama** Institute of Fluid Science, Tohoku University, Sendai, 98-8577, Japan *Present address,

More information

Automated Test Data Generation and Optimization Scheme Using Genetic Algorithm

Automated Test Data Generation and Optimization Scheme Using Genetic Algorithm 2011 International Conference on Software and Computer Applications IPCSIT vol.9 (2011) (2011) IACSIT Press, Singapore Automated Test Data Generation and Optimization Scheme Using Genetic Algorithm Roshni

More information

Application of a Genetic Algorithm to a Scheduling Assignement Problem

Application of a Genetic Algorithm to a Scheduling Assignement Problem Application of a Genetic Algorithm to a Scheduling Assignement Problem Amândio Marques a and Francisco Morgado b a CISUC - Center of Informatics and Systems of University of Coimbra, 3030 Coimbra, Portugal

More information

Applied Cloning Techniques for a Genetic Algorithm Used in Evolvable Hardware Design

Applied Cloning Techniques for a Genetic Algorithm Used in Evolvable Hardware Design Applied Cloning Techniques for a Genetic Algorithm Used in Evolvable Hardware Design Viet C. Trinh vtrinh@isl.ucf.edu Gregory A. Holifield greg.holifield@us.army.mil School of Electrical Engineering and

More information

Introduction to Genetic Algorithms

Introduction to Genetic Algorithms Advanced Topics in Image Analysis and Machine Learning Introduction to Genetic Algorithms Week 3 Faculty of Information Science and Engineering Ritsumeikan University Today s class outline Genetic Algorithms

More information

Evolved Multi-resolution Transforms for Optimized Image Compression and Reconstruction under Quantization

Evolved Multi-resolution Transforms for Optimized Image Compression and Reconstruction under Quantization Evolved Multi-resolution Transforms for Optimized Image Compression and Reconstruction under Quantization FRANK W. MOORE Mathematical Sciences Department University of Alaska Anchorage CAS 154, 3211 Providence

More information

International Conference on Computer Applications in Shipbuilding (ICCAS-2009) Shanghai, China Vol.2, pp

International Conference on Computer Applications in Shipbuilding (ICCAS-2009) Shanghai, China Vol.2, pp AUTOMATIC DESIGN FOR PIPE ARRANGEMENT CONSIDERING VALVE OPERATIONALITY H Kimura, Kyushu University, Japan S Iehira, Kyushu University, Japan SUMMARY We propose a novel evaluation method of valve operationality

More information

Multi-Objective Pipe Smoothing Genetic Algorithm For Water Distribution Network Design

Multi-Objective Pipe Smoothing Genetic Algorithm For Water Distribution Network Design City University of New York (CUNY) CUNY Academic Works International Conference on Hydroinformatics 8-1-2014 Multi-Objective Pipe Smoothing Genetic Algorithm For Water Distribution Network Design Matthew

More information

Multi-Objective Memetic Algorithm using Pattern Search Filter Methods

Multi-Objective Memetic Algorithm using Pattern Search Filter Methods Multi-Objective Memetic Algorithm using Pattern Search Filter Methods F. Mendes V. Sousa M.F.P. Costa A. Gaspar-Cunha IPC/I3N - Institute of Polymers and Composites, University of Minho Guimarães, Portugal

More information

An Evolutionary Algorithm Approach to Generate Distinct Sets of Non-Dominated Solutions for Wicked Problems

An Evolutionary Algorithm Approach to Generate Distinct Sets of Non-Dominated Solutions for Wicked Problems An Evolutionary Algorithm Approach to Generate Distinct Sets of Non-Dominated Solutions for Wicked Problems Marcio H. Giacomoni Assistant Professor Civil and Environmental Engineering February 6 th 7 Zechman,

More information