A HYBRID APPROACH TO GLOBAL OPTIMIZATION USING A CLUSTERING ALGORITHM IN A GENETIC SEARCH FRAMEWORK VIJAYKUMAR HANAGANDI. Los Alamos, NM 87545, U.S.A.

Size: px
Start display at page:

Download "A HYBRID APPROACH TO GLOBAL OPTIMIZATION USING A CLUSTERING ALGORITHM IN A GENETIC SEARCH FRAMEWORK VIJAYKUMAR HANAGANDI. Los Alamos, NM 87545, U.S.A."

Transcription

1 A HYBRID APPROACH TO GLOBAL OPTIMIZATION USING A CLUSTERING ALGORITHM IN A GENETIC SEARCH FRAMEWORK VIJAYKUMAR HANAGANDI MS C2, Los Alamos National Laboratory Los Alamos, NM 87545, U.S.A. and MICHAEL NIKOLAOU Chemical Engineering Department Texas A&M University College Station, TX , U.S.A. Submitted to Computers and Chemical Engineering, April 995. Abstract. The concern of this work is global optimization using genetic algorithms (GAs). In this work we propose a synergy between the cluster analysis technique, popular in classical stochastic global optimization, and the GA to accomplish global optimization. This synergy minimizes redundant searches around local optima and enhances the capability of the GA to explore new areas in the search space. The proposed methodology demonstrates superior performance when compared with the simple GA on benchmark cases. We also report our solution of the optimal pump congurations synthesis problem. *Author to whom all correspondence should be addressed. Tel.: (55) , FAX: (55) , vmh@lanl.gov

2 INTRODUCTION Multi-modal objective functions are common in engineering applications. Also, very little or no a priori information about the analytical properties of the objective function in some optimization applications in engineering stress the need to be able to search for the global optimum. Existing optimization techniques that seek the global optimum can be broadly classied into the following two categories: deterministic methods and stochastic methods. Here we focus on stochastic methods of optimization. Many classical stochastic optimization methods rely on the so called multistart algorithm. In this algorithm one starts with a set of uniformly random points and does local searches starting from a selected number of points from this set. A variety of algorithms stemmed out of the above basic scheme and some are very successful. Torn (978; 977) introduced an eective technique to prevent redundant objective function evaluations at local optima in a multistart algorithm. This technique uses a clustering algorithm. Experiences of various researchers (Kan and Timmer, 984; Boender et al., 982; Gomulka, 978a; 978b) on using Torn's clustering algorithm have shown that it is a valuable technique. The review paper by Kan and Timmer (984) discusses the salient features of several successful variations of the multistart algorithm and also gives a comparison between these techniques on some benchmark problems. The collection of papers edited by Dixon and Szego (978) is a citation classic in the eld of stochastic optimization. Among the non-traditional stochastic global optimization techniques, genetic algorithms (GAs) and simulated annealing (SA) have shown some promise. Genetic algorithms (GAs) are attracting a lot of attention in recent optimization literature. They are emerging as powerful alternatives to traditional optimization methods which are too restrictive and CPU intensive. A GA manipulates 2

3 populations of candidate solutions whereas a gradient-based method manipulates a single solution. This helps the GA to search for the global optimum. Also, since the search procedure is stochastic, the possibilities of a GA getting stuck in a local optimum are less. Moreover, GAs manipulate populations based solely on the individual objective function values and do not need any other information regarding the objective function. Theoretically, the simple genetic algorithm (SGA) locates the global optimum of a multi-modal function in nitely many generations (trials). In an SGA run, when a population is stuck in a local optimum, the mutation operator is supposed to provide the necessary diversity to the population to search in dierent areas. In spite of these arguments, it is demonstrated that an SGA in inecient (i.e. it takes a large number of trials) in locating the global optimum in a multi-modal function (Goldberg, 989) and that there is scope for improvement. De Jong (975) concluded that the elitist strategy (explained in the sequel) in an SGA brings premature convergence. In order to tackle this problem he proposed a modication that increases diversity in a population. His modication, called the crowding scheme, has seen some success. Androulakis and Vankatasubramanian (99) reported a modied genetic search which manipulates the search direction and called it an extended genetic search (EGS) algorithm. In this work we propose to use Torn's clustering algorithm in the genetic search procedure to identify and dissolve a niche (or a cluster) of species in a given population to increase population diversity. We demonstrate the eectiveness of our hybrid methodology on benchmark problems given in Dixon and Szego (978). Also, we report our solution of the multi-modal pump congurations design problem studied by Westerlund et al. (994). In the sections that follow, rst we give a background on genetic algorithms (GAs), then we discuss the motivation for our modication to the SGA. Subse-

4 quently we discuss methodology, simulation results and our solution of the optimal pump congurations synthesis problem using the modied GA. Conclusions appear in the last section. 2 BACKGROUND ON GENETIC ALGORITHMS GAs accomplish the task of optimization by starting with a random \population" of values for the parameters of an optimization problem, and thereafter producing new \generations" of improved values that combine the best \parts" of values from previous populations. In order to achieve this, it is very crucial to represent the objective variables in a non ambiguous way. In our case studies we use the binary representation of real numbers and integers and manipulate these strings to achieve reproduction, crossover and mutation. Parent representations which are used to produce ospring are selected in a random manner but with a bias toward selecting candidates with high tness values. This step is called reproduction. Ospring are produced by randomly cutting and swapping parts of the two parent representations. This is called the crossover operation. Mutation is achieved by screening the bits of the ospring and changing each binary bit to the opposite kind ( to or to ) with a very low probability. Mutation helps to introduce diversity in the population thus preventing the GA from getting saturated with solutions in local optima. Figs. and 2 illustrate the crossover and mutation operations using binary representations of the real and integer optimization variables. In each generation a limited number of candidates get to reproduce and have ospring which go into the next generation. When the problem is of objective function maximization the tness f i of candidate i in a given population is the normalized 4

5 Parent Representations Offspring Figure : Illustration of the crossover operation. The ospring are a result of cutting and swapping parts indicated between the arrows in the parent representations. Flip Figure 2: Illustration of mutation. Each bit is ipped to the opposite kind with a very low probability. 5

6 Table : The SGA pseudocode - Main algorithm. START generation count= initialize population by random strings repeat generation count=generation count+ create next generation until (generation count max gen) STOP objective function given by f i = f i? f min f max? f min where f i is the objective function value of candidate i and f max and f min are the maximum and minimum values respectively of f i encountered in that population. In a minimization problem the same procedure is followed but with a sign reversal. Elitism, in which the best candidate encountered thus far is reintroduced in each generation, can also be employed. Tables, 2 and give the pseudocode for a simple GA (SGA). In the pseudocode, the array tness contains the tness values of all the members of the current population. sumtness is the sum of the tness values of all members of the population and the function random() returns a uniform random number on the real interval [,]. One run of the roulette wheel selection routine gives two parents which can now be used to perform crossover and mutation as demonstrated in Figs. and 2. For further details about the theory and implementation of GAs we refer the reader to well known literature (Goldberg, 989; Davis, 99). The following points summarize the advantages of using GAs: As a GA proceeds randomly (yet systematically) in its search, it does not re- 6

7 Table 2: The SGA pseudocode - Renement of create next generation. START STOP j= repeat select two members by roulette wheel selection perform crossover perform mutation introduce ospring into the new population j=j+2 until(j population size) Table : The SGA pseudocode - Renement of select two members by roulette wheel selection. START STOP partial sum=. j= rand=random()sumtness repeat j=j+ partial sum=partial sum+tness(j) until(partial sum > rand or j=population size) selected individual=j 7

8 quire smoothness, derivability, continuity etc. in the objective function. The only requirement is that for a set of values for the optimization parameters, one must assign a tness value. GAs can eciently handle highly nonlinear and noisy objective functions encountered in stochastic processes where traditional gradient based methods are inecient. GAs are amenable to parallel processing. Unlike gradient search algorithms, in GA the objective function evaluation for one parameter set is independent of that for all others in the same generation. This facilitates the use of parallel computers for the search procedure. Motivation Many valuable lessons are learned by observing the nature closely. Holland (975) made arguments concerning the distribution of individuals in a natural environment. He argued that, in nature, when like individuals begin to dominate a niche, increased competition for limited resources decreases life expectancy and birth rates among the members of that niche. Less crowded niches experience less pressure and achieve life expectancy and birth rates much closer to their actual potential. In other words, each individual has a certain survival probability and reproduction capability associated with its tness in its environment. If a particular individual belongs to a niche which is dominated by like individuals then its survival and reproduction chances are diminished because it is more likely to compete with the members of its own community (niche) for limited resources whereas for an individual belonging to a less crowded niche these rates are not altered. To incorporate these aspects of nature in articial genetics, De Jong used 8

9 the so called crowding model. In this model, whenever an individual is born, it is compared with a xed number of individuals in the current population and it replaces an individual which is closest to itself. He called the xed number of individuals compared as the crowding factor. This helps prevent the duplication of like individuals and helps to maintain diversity to some extent. When multimodal function optimization is of concern, the issue of niches being formed has yet another dimension. Consider a situation where we start with a random initial population of parameters and do a few steps of gradient search starting from each one of these points. In this case there is a high probability that elements of a particular niche (which is nothing but a group of like individuals) tend toward the same local optimum if the gradient-based search is continued. Instead, it makes perfect sense to select the optimal member of this niche and do a gradient-based search starting from only this member. A similar concept was used by Torn (977; 978) in conjunction with the multistart algorithm- a stochastic optimization technique. Torn's algorithm, known as Algorithm LC (multiple local searches with clustering), is given below:. Choose uniformly random points. 2. Use a local search algorithm for a few steps.. Find clusters by using a cluster analysis technique. 4. Take a sample point from each cluster. Return to step 2. We propose to use this idea in the context of the genetic search procedure. We use Torn's clustering technique to identify and dissolve clusters in the parameter space so that redundant function evaluations of candidates leading to the same local 9

10 5 z 5 x Figure : The two-dimensional Shekel's function. y optimum are minimized and diversity is introduced in the population. In other words, by the cluster dissolving idea we explicitly impose the natural pressure on crowded niches and cause them to perish. When the performance of a simple genetic algorithm (SGA) and that of a simple genetic algorithm with De Jong's crowding model (SGA-CM) are compared we noticed that niches are not eectively dissolved for any value of the crowding factor. On the other hand our methodology explicitly identies and breaks the clusters. We use a two dimensional Shekel's function to demonstrate this. This test function has six prominent local maxima. Details of this function are given in the next subsection. Fig. graphically shows this function. Fig. 4 shows a snap shot of the parameter values in the twentieth generation of an SGA. Clusters are pointed out by arrows. Fig. 5 shows the corresponding snapshot of an SGA-CM run. Note that the clusters are still present. Similar behavior is noted for several generations and during several dierent runs. Fig. 6 shows that the GA with a clustering algorithm (SGA-CL) eectively dissolves the clusters. In our method, as soon as niches are

11 8 6 x x Figure 4: The distribution of the variables after 2 generations in an SGA run for the 2-D Shekel's function. 8 6 x x Figure 5: The distribution of the variables after 2 generations in an SGA-CM run for the 2-D Shekel's function.

12 8 x x Figure 6: The distribution of the variables after 2 generations in an SGA-CL run for the 2-D Shekel's function. formed they are identied. The best members of those niches are retained and the rest are randomized. This introduces diversity in the population as shown in the spread of the parameters in Fig. 6. We show the same aspects again with the Goldstein and Price function in Figs. 7, 8, 9 and. 4 Methodology and Simulation Results Although it eliminates redundant calculations, Torn's algorithm (see previous section) lacks the ability to search in new directions which is crucial for multi-modal function optimization. To remedy this problem we propose a hybrid methodology which relies on the strength of GAs to search in new directions and on the ability of the clustering technique to reduce redundant calculations. The steps in the proposed GA with clustering (GA-CL) are as follows:. Start with an initial population. 2. Evaluate the tness of each member of the population. 2

13 z x y Figure 7: The Goldstein and Price function. 2 x x Figure 8: The distribution of the variables after 2 generations in an SGA run for the Goldstein and Price function.

14 2.5 x s R x Figure 9: The distribution of the variables after 2 generations in an SGA-CM run for the Goldstein and Price function. x x Figure : The distribution of the variables after 2 generations in an SGA-CL run for the Goldstein and Price function. 4

15 . Apply a clustering algorithm to identify clusters. 4. If any clusters are identied, from each cluster select the best member and randomize the rest. 5. Evaluate tness of the new members whose tness has not been evaluated in step Apply reproduction, crossover and mutation. 7. Generate new population. 8. Stop if the number of generations exceeds the maximum allowable generations or if convergence is satisfactory. Else, go to step 2. The breaking of clusters reduces redundant objective function evaluations and the randomization introduces diversity in the population. In the GA we use here, we employ elitism. Elitism accelerates local search. Using elitism in between cluster dissolving operations is analogous to doing a local search between cluster dissolving operations as suggested by Torn in his algorithm LC. The clustering algorithm we use here aims at selecting a seed point and growing the cluster by enlarging the hypersphere around the seed point as long as the density of the cluster remains greater than the average density of the space in which all the points are distributed. This method does not require the user to specify the number of clusters a priori. Details of this algorithm are given in APPENDIX A. We report our study on six benchmark problems and one design problem. The design problem is given in the next section. The benchmark problems are taken from Dixon and Szego (978). Details of these problems are given in Dixon and Szego (978). Table 4 summarizes the properties of these problems. Table 5 gives a comparison between the success probabilities of dierent approaches. These 5

16 Table 4: Details of the benchmark optimization problems GP, H, H6, S5, S7 and S. Goldstein and Price (GP): max f? [( + (x + x 2 + ) 2 (9? 4x + x 2? 4x 2 + 6x x 2 + x 2 2] [ + (2x? x 2 ) 2 (8? 2x + 2x x 2? 6x x x 2 2 ]g Range:?2: x ;2 2: Four local maxima. Global maximum at: x = :, x 2 =?:7 with the value -. Hartman's Family (H: variables, H6: 6 variables) max f P m i= c iexp(? P n j= a ij(x j? p ij ) 2 g, n= or n=6 Range: : x i : Four local maxima. Global maxima at: H:(.,.556,.85) with the value.86 H6:(.29,.42,.47,.265,.9,.66) with the value.4 Shekel's Family (S5: m=5, S7: m=7, S: m=): max f P m i= Pn Range: : x i : Global maxima at: j= (x i?a ij )2 +ci g, n=4 S5: (4.,.998,.99,.996) with the value.422 S7: (4.9, 4., 4., 4.7) with the value.99 S: (.995, 4.5, 4.,.996) with the value.5 6

17 Table 5: Comparison between success probabilities of dierent algorithms. Algorithms Problem SGA SGA-CM EGS SGA-CL Goldstein and Price (GP) Hartman (H) Hartman 6 (H6) Shekel 5 (S5) Shekel 7 (S7) Shekel (S) probabilities were computed after one hundred independent runs for each problem. The parameters for the various algorithms compared are given in Table 6. The parameters were adjusted by trial and error to get the best performance of the SGA and SGA-CM approaches. The simple genetic algorithm (SGA) gives the base case for comparison. De Jong's crowding model (SGA-CM) shows marginal to no improvement over the performance of the SGA. The SGA with clustering approach (SGA-CL) showed superior performance in all cases. Figs. through 6 show the convergence proles of SGA and SGA-CL approaches for dierent problems. The dotted lines on all these plots indicate the global optimum of the function being investigated. Since a genetic search-based algorithm is a stochastic search, a single run does not give the true convergence prole. To present a reasonably true prole the convergence plots are made by averaging fty independent runs. The improvement of the clustering approach over the SGA is clear from these plots too. In terms of objective function computations, the SGA tops the list as it is based on non overlapping populations (in the transition from one generation to the next, the whole population is replaced by the ospring.) 7

18 Table 6: GA parameters. Population size : Number of ospring in each generation : for SGA, 2 for SGA-CM and 2 for SGA-CL Mutation rate :. Crossover probability :. Elitism : O for SGA, O for SGA-CM and On for SGA-CL Crowding factor : (only applicable for SGA-CM) Chromosome length per variable : 5-5 SGA-CL F(x) SGA Generation Figure : The objective function vs. generation number for GP (average of fty independent runs). 8

19 SGA-CL.855 F(x) SGA Generations Figure 2: The objective function vs. generation number for H (average of fty independent runs) SGA-CL.24 F(x).22.2 SGA Generations Figure : The objective function vs. generation number for H6 (average of fty independent runs). 9

20 9 SGA-CL 8 F(x) SGA Generations Figure 4: The objective function vs. generation number for S5 (average of fty independent runs). 9 SGA-CL F(x) SGA Generations Figure 5: The objective function vs. generation number for S7 (average of fty independent runs). 2

21 9 SGA-CL F(x) SGA Generations Figure 6: The objective function vs. generation number for S (average of fty independent runs). 5 Optimization of Pump Congurations as an MINLP Problem In this section we solve the problem of optimal pump conguration synthesis (Westerlund et al., 994). The optimization problem is cast as a constrained MINLP problem which is outlined next. Centrifugal pumps are available at variable speeds (rpm). The total required pressure rise and the total ow are specied. The aim is to select the best seriesparallel combination of the pumps that minimizes total cost. The cost is a function of the number of parallel pumping lines (N p;i ) in each level, the number of pumps (N s;i ) in series in every line and the power (P i ) of each pump. The power required for each pump is a function of the pump speed (! i ) and the volumetric ow in the pump ( V _ i ). Fig. 7 gives a schematic of a two level pump conguration. Problem details are given in Table 7. We use the three level pump conguration (L = ) in our application. Westerlund et al. (994) have reported that this is a nonconvex 2

22 Table 7: Details of the pump congurations problem. Objective Variables N p;i number of parallel pumping lines N s;i number of pumps in each line P i power of pump i x i volumetric fraction of Vtot _ in i! i rpm of pump i Constraints p = 629(! 295 )2 + :696(! ) V _ 295? :6 V _ 2 p 2 = 25(! )2 + 2:95(! 2 ) V _ 295 2? :5 V _ 2 2 p = 6(! 295 )2 + :5(! ) V _ 295? :946 V _ 2! i 295 P Objective Function L J = i=(c i + Ci P i)n p;i N s;i where L =, Ci = 8, C = 629: C 2 = 2489:, C = 67:89, _V tot = 5, p tot = 4 P = (! )(9:9(! )2 + :6(! ) V _? :56 V_ P 2 = (! 2 )(:2(! )2 + :644(! 2 ) V _ 295 2? :564 _ P = (! )(6:52(! )2 + :2(! ) V _ 295 2? :22 V _ 2 _V i = x i Vtot _ p i = N p;i ptot N s;i ) V 2 2 ) ) 22

23 Level 2 Level L=2, N =2, i=,2 pi N =2, i=,2 si Figure 7: Schematic of a two level pump conguration. Each level is shown with two parallel lines with three pumps in each line. problem with thirty seven local minima. They used the extended cutting plane (ECP) method and the software DICOPT++ to solve the MINLP problem. They had to change the problem formulation to improve the solution and achieve near global optimum results. Also, proper selection of the software related parameters was found to be critical. We found that the GA-CL could solve this MINLP problem and it reached the global optimum in a reasonable number of generations. The objective function augmented with the weighted constraints was used to evaluate tness of each candidate solution. Fig. 8 shows the convergence prole. Fig. 9 gives the globally optimal pump conguration. Table 8 gives a comparison of the dierent solutions. An advantage of using GAs for optimization is that in the last population, a variety of near optimal solutions are present from which the designer can choose in case the optimal solution has some undesirable properties or if the suboptimal solution has some additional desirable properties. In one of the suboptimal designs 2

24 -5 -e+6 -.5e+6?J -2e+6-2.5e+6 -e+6 -.5e Generation Figure 8:.-J vs. generation for the optimal pump congurations synthesis problem. Level 2 Level Figure 9: The globally optimal pump conguration. 24

25 Table 8: Comparison of results for the pump congurations problem. Design# Conguration Cost Remarks N p; = 2, N p;2 =, N p; = Global N s; =, N s;2 = 2, N s; = optimum x = :94, x 2 = :86, x = : (Westerlund et al.,! = 2855,! 2 = 295,! = 994) 2 N p; =, N p;2 =, N p; = 2 54 ECP method N s; =, N s;2 =, N s; = 2 solution x = :449, x 2 = :, x = :55 (Westerlund et al.,! = 286,! 2 =,! = ) N p; =, N p;2 =, N p; = 557 DICOPT++ N s; =, N s;2 =, N s; = 2 solution x = :, x 2 = :, x = (Westerlund et al.,! =,! 2 =,! = ) 4 N p; = 2, N p;2 =, N p; = GA solution N s; =, N s;2 = 2, N s; = x = :94, x 2 = :86, x = :! = 2855,! 2 = 295,! = 5 N p; = 2, N p;2 =, N p; = 2578 GA suboptimal N s; =, N s;2 =, N s; = solution. x = :, x 2 = :, x = : Note that!! = 296,! 2 =,! = violates the maximum rpm limit for a pump 25

26 Level Figure 2: The suboptimal pump conguration. (design#5 in Table 8) the GA suggests that if pumps of 296 rpm ( rpm more than the maximum limit) are used we can get a better design consisting of only one level with two parallel lines containing one pump in each line as shown in Fig Conclusions In stochastic search techniques like GAs, formation of niches around local optima is detrimental to the convergence to global optimum. In this paper we presented a modication to the simple genetic search which is based on a clustering algorithm and demonstrated its eciency in solving benchmark multi-modal optimization problems. This algorithm aims at identifying and dissolving the clusters so that diversity is preserved in the population. The superiority of the proposed methodology over the simple genetic search and extended genetic search is clear. Our methodology was able to nd the globally optimal conguration in the optimal pump congurations synthesis problem. In this problem, apart from the glob- 26

27 ally optimal conguration, the modied GA was also able to locate a suboptimal, simple and trivial conguration. 27

28 References Androulakis, I. P., and V. Venkatasubramanian, \A Genetic Algorithmic Framework for Process Design and Optimization," Computers and Chem. Eng., 5, 27 (99). Boender, C. G. E., A. H. G. R. Kan, and G. T. Timmer, \A Stochastic Method for Global Optimization," Mathematical Programming, 22, 25 (982). Davis, L, 99, Handbook of Genetic Algorithms, Van Nostrand, New York. De Jong, K. A., An Analysis of the Behavior of a Class of Genetic Adaptive Systems, Ph.D. Dissertation, University of Michigan, Ann Arbor, MI (975). Dixon, L. C. W., and G. P. Szego, Towards Global Optimisation 2, North-Holland, Amsterdam (978). Goldberg, D. E., 989, Genetic Algorithms in Search Optimization and Machine Learning, Addison Wesley, Reading, MA. Gomulka, J., \Deterministic Versus Probabilistic Approaches to Global Optimization," in Towards Global Optimisation 2, L. C. W. Dixon and G. P. Szego (Eds.), North-Holland, Amsterdam (978a). Gomulka, J., \A User's Experience with Torn's Algorithm," in Towards Global Optimisation 2, L. C. W. Dixon and G. P. Szego (Eds.), North-Holland, Amsterdam (978b). Holland, J. H., Adaptation in Natural and Articial Systems, The University of Michigan Press, Ann Arbor, MI (975). 28

29 Kan, A. H. G. R., and G. T. Timmer, \A Stochastic Approach to Global Optimization," in Numerical Optimization, SIAM, Philadelphia, PA (984). More, J. J., and S. J. Wright, Optimization Software Guide, SIAM, Philadelphia, PA (99). Torn, A. A., \A Search-Clustering Approach to Global Optimization," in Towards Global Optimisation 2, L. C. W. Dixon and G. P. Szego (Eds.), North- Holland, Amsterdam (978). Torn, A. A., \Cluster Analysis Using Seed Points and Density-Determined Hyperspheres as an Aid to Global Optimization," IEEE Trans. Sys. Man. Cyber., 7, 6 (977). Westerlund, T., F. Pettersson, and I. E. Grossmann, \Optimization of Pump Congurations as a MINLP Problem," Computers and Chem. Eng., 8, 845 (994). 29

30 APPENDIX A T ORN'S CLUSTERING ALGORITHM In what follows we describe briey the clustering algorithm used by Torn in his algorithm LC. The basic idea is to grow clusters from seed points to include all the neighboring points. There are two main issues in identifying clusters: Location of the seed points. When to stop the growth of a cluster. In other words, deciding about a threshold distance r for stopping the growth of a cluster. The choice of a threshold distance is critical because if r is large, all points will be assigned to a single cluster and if r is small, each point will be form an isolated cluster. In a multi-modal objective function minimization problem, consider a collection of N points in which we want to identify clusters. The assumption is made that the local minima would be good seed points to use and that the point with the smallest function value may be used as an estimate of the local minimum. The cluster growing is started from this point. When this cluster stops growing and a new seed point is needed, the point, among those not yet classied, with the least function value is chosen. This procedure is repeated until all the N points are classied. A cluster is grown by enlarging a hypersphere around the seed point as long as the point density in the hypersphere remains greater than the uniform point density = N V where V is the volume of the space containing all the N points and is given by V = N i=4 2 i :

31 i are the eigenvalues of the covariance matrix S and s ij = N? ( NX k= x ik x jk? n nx nx x ik k= k= In the above equation, x k is the kth parameter vector and x ik is the ith component of x k. Torn's original clustering algorithm includes only the parameters of optimization in the vectors x k for cluster identication, but we include the objective function value along with the parameters of optimization in the vectors x k. This modication gives better classication (cluster identication) particularly for functions with multiple local minima, each having dierent objective function values, which are crowded in a particular search space. x jk )

32 List of Figures Illustration of the crossover operation. The ospring are a result of cutting and swapping parts indicated between the arrows in the parent representations. : : : : : : : : : : : : : : : : : : : : : : : : 5 2 Illustration of mutation. Each bit is ipped to the opposite kind with a very low probability. : : : : : : : : : : : : : : : : : : : : : 5 The two-dimensional Shekel's function. : : : : : : : : : : : : : : : 4 The distribution of the variables after 2 generations in an SGA run for the 2-D Shekel's function. : : : : : : : : : : : : : : : : : : 5 The distribution of the variables after 2 generations in an SGA- CM run for the 2-D Shekel's function. : : : : : : : : : : : : : : : : 6 The distribution of the variables after 2 generations in an SGA-CL run for the 2-D Shekel's function. : : : : : : : : : : : : : : : : : : 2 7 The Goldstein and Price function. : : : : : : : : : : : : : : : : : : 8 The distribution of the variables after 2 generations in an SGA run for the Goldstein and Price function. : : : : : : : : : : : : : : 4 9 The distribution of the variables after 2 generations in an SGA- CM run for the Goldstein and Price function. : : : : : : : : : : : 4 The distribution of the variables after 2 generations in an SGA-CL run for the Goldstein and Price function. : : : : : : : : : : : : : : 5 The objective function vs. generation number for GP (average of fty independent runs). : : : : : : : : : : : : : : : : : : : : : : : : 9 2 The objective function vs. generation number for H (average of fty independent runs). : : : : : : : : : : : : : : : : : : : : : : : : 9 2

33 The objective function vs. generation number for H6 (average of fty independent runs). : : : : : : : : : : : : : : : : : : : : : : : : 2 4 The objective function vs. generation number for S5 (average of fty independent runs). : : : : : : : : : : : : : : : : : : : : : : : : 2 5 The objective function vs. generation number for S7 (average of fty independent runs). : : : : : : : : : : : : : : : : : : : : : : : : 2 6 The objective function vs. generation number for S (average of fty independent runs). : : : : : : : : : : : : : : : : : : : : : : : : 2 7 Schematic of a two level pump conguration. Each level is shown with two parallel lines with three pumps in each line. : : : : : : : 2 8.-J vs. generation for the optimal pump congurations synthesis problem. : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 2 9 The globally optimal pump conguration. : : : : : : : : : : : : : 25 2 The suboptimal pump conguration. : : : : : : : : : : : : : : : : 27

34 List of Tables The SGA pseudocode - Main algorithm. : : : : : : : : : : : : : : 6 2 The SGA pseudocode - Renement of create next generation. : : : 7 The SGA pseudocode - Renement of select two members by roulette wheel selection. : : : : : : : : : : : : : : : : : : : : : : : : : : : : 7 4 Details of the benchmark optimization problems GP, H, H6, S5, S7 and S. : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 7 5 Comparison between success probabilities of dierent algorithms. 8 6 GA parameters. : : : : : : : : : : : : : : : : : : : : : : : : : : : : 8 7 Details of the pump congurations problem. : : : : : : : : : : : : 24 8 Comparison of results for the pump congurations problem. : : : 26 4

SOLUTION OF THE OPTIMAL PLANT LOCATION AND SIZING PROBLEM USING. and. Michael Nikolaou. Texas A&M University

SOLUTION OF THE OPTIMAL PLANT LOCATION AND SIZING PROBLEM USING. and. Michael Nikolaou. Texas A&M University SOLUTION OF THE OPTIMAL PLANT LOCATION AND SIZING PROBLEM USING SIMULATED ANNEALING AND GENETIC ALGORITHMS Ramana Rao, Vijaykumar Hanagandi and Kevin Buescher MS C32, Los Alamos National Laboratory Los

More information

V.Petridis, S. Kazarlis and A. Papaikonomou

V.Petridis, S. Kazarlis and A. Papaikonomou Proceedings of IJCNN 93, p.p. 276-279, Oct. 993, Nagoya, Japan. A GENETIC ALGORITHM FOR TRAINING RECURRENT NEURAL NETWORKS V.Petridis, S. Kazarlis and A. Papaikonomou Dept. of Electrical Eng. Faculty of

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

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

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

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

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

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

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

Adaptive Elitist-Population Based Genetic Algorithm for Multimodal Function Optimization

Adaptive Elitist-Population Based Genetic Algorithm for Multimodal Function Optimization Adaptive Elitist-Population Based Genetic Algorithm for Multimodal Function ptimization Kwong-Sak Leung and Yong Liang Department of Computer Science & Engineering, The Chinese University of Hong Kong,

More information

Binary Representations of Integers and the Performance of Selectorecombinative Genetic Algorithms

Binary Representations of Integers and the Performance of Selectorecombinative Genetic Algorithms Binary Representations of Integers and the Performance of Selectorecombinative Genetic Algorithms Franz Rothlauf Department of Information Systems University of Bayreuth, Germany franz.rothlauf@uni-bayreuth.de

More information

Frontier Pareto-optimum

Frontier Pareto-optimum Distributed Genetic Algorithms with a New Sharing Approach in Multiobjective Optimization Problems Tomoyuki HIROYASU Mitsunori MIKI Sinya WATANABE Doshisha University, Dept. of Knowledge Engineering and

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

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

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

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

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

A Population-Based Learning Algorithm Which Learns Both. Architectures and Weights of Neural Networks y. Yong Liu and Xin Yao

A Population-Based Learning Algorithm Which Learns Both. Architectures and Weights of Neural Networks y. Yong Liu and Xin Yao A Population-Based Learning Algorithm Which Learns Both Architectures and Weights of Neural Networks y Yong Liu and Xin Yao Computational Intelligence Group Department of Computer Science University College,

More information

A genetic algorithm implemented in Matlab is presented. Matlab is used for the following reasons:

A genetic algorithm implemented in Matlab is presented. Matlab is used for the following reasons: A Genetic Algorithm for Function Optimization: A Matlab Implementation Christopher R. Houck North Carolina State University and Jeery A. Joines North Carolina State University and Michael G. Kay North

More information

336 THE STATISTICAL SOFTWARE NEWSLETTER where z is one (randomly taken) pole of the simplex S, g the centroid of the remaining d poles of the simplex

336 THE STATISTICAL SOFTWARE NEWSLETTER where z is one (randomly taken) pole of the simplex S, g the centroid of the remaining d poles of the simplex THE STATISTICAL SOFTWARE NEWSLETTER 335 Simple Evolutionary Heuristics for Global Optimization Josef Tvrdk and Ivan Krivy University of Ostrava, Brafova 7, 701 03 Ostrava, Czech Republic Phone: +420.69.6160

More information

Exploration vs. Exploitation in Differential Evolution

Exploration vs. Exploitation in Differential Evolution Exploration vs. Exploitation in Differential Evolution Ângela A. R. Sá 1, Adriano O. Andrade 1, Alcimar B. Soares 1 and Slawomir J. Nasuto 2 Abstract. Differential Evolution (DE) is a tool for efficient

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

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

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

x GP Function 6 x 2 x 1

x GP Function 6 x 2 x 1 DIRECT Optimization Algorithm User Guide Daniel E. Finkel Center for Research in Scientific Computation North Carolina State University Raleigh, NC 27695-8205 definkel@unity.ncsu.edu March 2, 2003 Abstract

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

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

AN EVOLUTIONARY APPROACH TO DISTANCE VECTOR ROUTING

AN EVOLUTIONARY APPROACH TO DISTANCE VECTOR ROUTING International Journal of Latest Research in Science and Technology Volume 3, Issue 3: Page No. 201-205, May-June 2014 http://www.mnkjournals.com/ijlrst.htm ISSN (Online):2278-5299 AN EVOLUTIONARY APPROACH

More information

Genetic Algorithms For Vertex. Splitting in DAGs 1

Genetic Algorithms For Vertex. Splitting in DAGs 1 Genetic Algorithms For Vertex Splitting in DAGs 1 Matthias Mayer 2 and Fikret Ercal 3 CSC-93-02 Fri Jan 29 1993 Department of Computer Science University of Missouri-Rolla Rolla, MO 65401, U.S.A. (314)

More information

Reducing Graphic Conflict In Scale Reduced Maps Using A Genetic Algorithm

Reducing Graphic Conflict In Scale Reduced Maps Using A Genetic Algorithm Reducing Graphic Conflict In Scale Reduced Maps Using A Genetic Algorithm Dr. Ian D. Wilson School of Technology, University of Glamorgan, Pontypridd CF37 1DL, UK Dr. J. Mark Ware School of Computing,

More information

Adaptive Crossover in Genetic Algorithms Using Statistics Mechanism

Adaptive Crossover in Genetic Algorithms Using Statistics Mechanism in Artificial Life VIII, Standish, Abbass, Bedau (eds)(mit Press) 2002. pp 182 185 1 Adaptive Crossover in Genetic Algorithms Using Statistics Mechanism Shengxiang Yang Department of Mathematics and Computer

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

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

Artificial Intelligence Application (Genetic Algorithm)

Artificial Intelligence Application (Genetic Algorithm) Babylon University College of Information Technology Software Department Artificial Intelligence Application (Genetic Algorithm) By Dr. Asaad Sabah Hadi 2014-2015 EVOLUTIONARY ALGORITHM The main idea about

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

Information Fusion Dr. B. K. Panigrahi

Information Fusion Dr. B. K. Panigrahi Information Fusion By Dr. B. K. Panigrahi Asst. Professor Department of Electrical Engineering IIT Delhi, New Delhi-110016 01/12/2007 1 Introduction Classification OUTLINE K-fold cross Validation Feature

More information

CHAPTER 4 GENETIC ALGORITHM

CHAPTER 4 GENETIC ALGORITHM 69 CHAPTER 4 GENETIC ALGORITHM 4.1 INTRODUCTION Genetic Algorithms (GAs) were first proposed by John Holland (Holland 1975) whose ideas were applied and expanded on by Goldberg (Goldberg 1989). GAs is

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

Genetic Algorithms for Real Parameter Optimization

Genetic Algorithms for Real Parameter Optimization Genetic Algorithms for Real Parameter Optimization Alden H. Wright Department of Computer Science University of Montana Missoula, Montana 59812 Abstract This paper is concerned with the application of

More information

Department of Electrical Engineering, Keio University Hiyoshi Kouhoku-ku Yokohama 223, Japan

Department of Electrical Engineering, Keio University Hiyoshi Kouhoku-ku Yokohama 223, Japan Shape Modeling from Multiple View Images Using GAs Satoshi KIRIHARA and Hideo SAITO Department of Electrical Engineering, Keio University 3-14-1 Hiyoshi Kouhoku-ku Yokohama 223, Japan TEL +81-45-563-1141

More information

Learning Adaptive Parameters with Restricted Genetic Optimization Method

Learning Adaptive Parameters with Restricted Genetic Optimization Method Learning Adaptive Parameters with Restricted Genetic Optimization Method Santiago Garrido and Luis Moreno Universidad Carlos III de Madrid, Leganés 28911, Madrid (Spain) Abstract. Mechanisms for adapting

More information

CS5401 FS2015 Exam 1 Key

CS5401 FS2015 Exam 1 Key CS5401 FS2015 Exam 1 Key This is a closed-book, closed-notes exam. The only items you are allowed to use are writing implements. Mark each sheet of paper you use with your name and the string cs5401fs2015

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

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

Review: Final Exam CPSC Artificial Intelligence Michael M. Richter

Review: Final Exam CPSC Artificial Intelligence Michael M. Richter Review: Final Exam Model for a Learning Step Learner initially Environm ent Teacher Compare s pe c ia l Information Control Correct Learning criteria Feedback changed Learner after Learning Learning by

More information

A Novel Approach to Planar Mechanism Synthesis Using HEEDS

A Novel Approach to Planar Mechanism Synthesis Using HEEDS AB-2033 Rev. 04.10 A Novel Approach to Planar Mechanism Synthesis Using HEEDS John Oliva and Erik Goodman Michigan State University Introduction The problem of mechanism synthesis (or design) is deceptively

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

A Genetic Algorithm-Based Approach for Energy- Efficient Clustering of Wireless Sensor Networks

A Genetic Algorithm-Based Approach for Energy- Efficient Clustering of Wireless Sensor Networks A Genetic Algorithm-Based Approach for Energy- Efficient Clustering of Wireless Sensor Networks A. Zahmatkesh and M. H. Yaghmaee Abstract In this paper, we propose a Genetic Algorithm (GA) to optimize

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

[Premalatha, 4(5): May, 2015] ISSN: (I2OR), Publication Impact Factor: (ISRA), Journal Impact Factor: 2.114

[Premalatha, 4(5): May, 2015] ISSN: (I2OR), Publication Impact Factor: (ISRA), Journal Impact Factor: 2.114 IJESRT INTERNATIONAL JOURNAL OF ENGINEERING SCIENCES & RESEARCH TECHNOLOGY GENETIC ALGORITHM FOR OPTIMIZATION PROBLEMS C. Premalatha Assistant Professor, Department of Information Technology Sri Ramakrishna

More information

Module 1 Lecture Notes 2. Optimization Problem and Model Formulation

Module 1 Lecture Notes 2. Optimization Problem and Model Formulation Optimization Methods: Introduction and Basic concepts 1 Module 1 Lecture Notes 2 Optimization Problem and Model Formulation Introduction In the previous lecture we studied the evolution of optimization

More information

Active contour: a parallel genetic algorithm approach

Active contour: a parallel genetic algorithm approach id-1 Active contour: a parallel genetic algorithm approach Florence Kussener 1 1 MathWorks, 2 rue de Paris 92196 Meudon Cedex, France Florence.Kussener@mathworks.fr Abstract This paper presents an algorithm

More information

OPTIMAL DESIGN OF WATER DISTRIBUTION SYSTEMS BY A COMBINATION OF STOCHASTIC ALGORITHMS AND MATHEMATICAL PROGRAMMING

OPTIMAL DESIGN OF WATER DISTRIBUTION SYSTEMS BY A COMBINATION OF STOCHASTIC ALGORITHMS AND MATHEMATICAL PROGRAMMING 2008/4 PAGES 1 7 RECEIVED 18. 5. 2008 ACCEPTED 4. 11. 2008 M. ČISTÝ, Z. BAJTEK OPTIMAL DESIGN OF WATER DISTRIBUTION SYSTEMS BY A COMBINATION OF STOCHASTIC ALGORITHMS AND MATHEMATICAL PROGRAMMING ABSTRACT

More information

Computational Intelligence

Computational Intelligence Computational Intelligence Module 6 Evolutionary Computation Ajith Abraham Ph.D. Q What is the most powerful problem solver in the Universe? ΑThe (human) brain that created the wheel, New York, wars and

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

Introduction to Genetic Algorithms. Based on Chapter 10 of Marsland Chapter 9 of Mitchell

Introduction to Genetic Algorithms. Based on Chapter 10 of Marsland Chapter 9 of Mitchell Introduction to Genetic Algorithms Based on Chapter 10 of Marsland Chapter 9 of Mitchell Genetic Algorithms - History Pioneered by John Holland in the 1970s Became popular in the late 1980s Based on ideas

More information

GENETIC ALGORITHM VERSUS PARTICLE SWARM OPTIMIZATION IN N-QUEEN PROBLEM

GENETIC ALGORITHM VERSUS PARTICLE SWARM OPTIMIZATION IN N-QUEEN PROBLEM Journal of Al-Nahrain University Vol.10(2), December, 2007, pp.172-177 Science GENETIC ALGORITHM VERSUS PARTICLE SWARM OPTIMIZATION IN N-QUEEN PROBLEM * Azhar W. Hammad, ** Dr. Ban N. Thannoon Al-Nahrain

More information

Numerical Experiments with a Population Shrinking Strategy within a Electromagnetism-like Algorithm

Numerical Experiments with a Population Shrinking Strategy within a Electromagnetism-like Algorithm Numerical Experiments with a Population Shrinking Strategy within a Electromagnetism-like Algorithm Ana Maria A. C. Rocha and Edite M. G. P. Fernandes Abstract This paper extends our previous work done

More information

An Evolutionary Algorithm for Minimizing Multimodal Functions

An Evolutionary Algorithm for Minimizing Multimodal Functions An Evolutionary Algorithm for Minimizing Multimodal Functions D.G. Sotiropoulos, V.P. Plagianakos and M.N. Vrahatis University of Patras, Department of Mamatics, Division of Computational Mamatics & Informatics,

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

Escaping Local Optima: Genetic Algorithm

Escaping Local Optima: Genetic Algorithm Artificial Intelligence Escaping Local Optima: Genetic Algorithm Dae-Won Kim School of Computer Science & Engineering Chung-Ang University We re trying to escape local optima To achieve this, we have learned

More information

An evolutionary annealing-simplex algorithm for global optimisation of water resource systems

An evolutionary annealing-simplex algorithm for global optimisation of water resource systems FIFTH INTERNATIONAL CONFERENCE ON HYDROINFORMATICS 1-5 July 2002, Cardiff, UK C05 - Evolutionary algorithms in hydroinformatics An evolutionary annealing-simplex algorithm for global optimisation of water

More information

Gradient Controlled, Typical Distance Clustering for Global Optimization.

Gradient Controlled, Typical Distance Clustering for Global Optimization. Gradient Controlled, Typical Distance Clustering for Global Optimization. I. G. Tsoulos and I. E. Lagaris Department of Computer Science, University of Ioannina P.O.Box 1186, Ioannina 45110 - GREECE Abstract

More information

3 Linkage identication. In each linkage. Intra GA Intra GA Intra GA. BB candidates. Inter GA;

3 Linkage identication. In each linkage. Intra GA Intra GA Intra GA. BB candidates. Inter GA; Designing a Genetic Algorithm Using the Linkage Identication by Nonlinearity Check Masaharu Munetomo and David E. Goldberg IlliGAL Report No. 98014 December 1998 Illinois Genetic Algorithms Laboratory

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

1. Introduction. 2. Motivation and Problem Definition. Volume 8 Issue 2, February Susmita Mohapatra

1. Introduction. 2. Motivation and Problem Definition. Volume 8 Issue 2, February Susmita Mohapatra Pattern Recall Analysis of the Hopfield Neural Network with a Genetic Algorithm Susmita Mohapatra Department of Computer Science, Utkal University, India Abstract: This paper is focused on the implementation

More information

Crew Scheduling Problem: A Column Generation Approach Improved by a Genetic Algorithm. Santos and Mateus (2007)

Crew Scheduling Problem: A Column Generation Approach Improved by a Genetic Algorithm. Santos and Mateus (2007) In the name of God Crew Scheduling Problem: A Column Generation Approach Improved by a Genetic Algorithm Spring 2009 Instructor: Dr. Masoud Yaghini Outlines Problem Definition Modeling As A Set Partitioning

More information

An Efficient Constraint Handling Method for Genetic Algorithms

An Efficient Constraint Handling Method for Genetic Algorithms An Efficient Constraint Handling Method for Genetic Algorithms Kalyanmoy Deb Kanpur Genetic Algorithms Laboratory (KanGAL) Department of Mechanical Engineering Indian Institute of Technology Kanpur Kanpur,

More information

Object Modeling from Multiple Images Using Genetic Algorithms. Hideo SAITO and Masayuki MORI. Department of Electrical Engineering, Keio University

Object Modeling from Multiple Images Using Genetic Algorithms. Hideo SAITO and Masayuki MORI. Department of Electrical Engineering, Keio University Object Modeling from Multiple Images Using Genetic Algorithms Hideo SAITO and Masayuki MORI Department of Electrical Engineering, Keio University E-mail: saito@ozawa.elec.keio.ac.jp Abstract This paper

More information

An algorithm for censored quantile regressions. Abstract

An algorithm for censored quantile regressions. Abstract An algorithm for censored quantile regressions Thanasis Stengos University of Guelph Dianqin Wang University of Guelph Abstract In this paper, we present an algorithm for Censored Quantile Regression (CQR)

More information

Monika Maharishi Dayanand University Rohtak

Monika Maharishi Dayanand University Rohtak Performance enhancement for Text Data Mining using k means clustering based genetic optimization (KMGO) Monika Maharishi Dayanand University Rohtak ABSTRACT For discovering hidden patterns and structures

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

Hyperplane Ranking in. Simple Genetic Algorithms. D. Whitley, K. Mathias, and L. Pyeatt. Department of Computer Science. Colorado State University

Hyperplane Ranking in. Simple Genetic Algorithms. D. Whitley, K. Mathias, and L. Pyeatt. Department of Computer Science. Colorado State University Hyperplane Ranking in Simple Genetic Algorithms D. Whitley, K. Mathias, and L. yeatt Department of Computer Science Colorado State University Fort Collins, Colorado 8523 USA whitley,mathiask,pyeatt@cs.colostate.edu

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

ARTIFICIAL INTELLIGENCE (CSCU9YE ) LECTURE 5: EVOLUTIONARY ALGORITHMS

ARTIFICIAL INTELLIGENCE (CSCU9YE ) LECTURE 5: EVOLUTIONARY ALGORITHMS ARTIFICIAL INTELLIGENCE (CSCU9YE ) LECTURE 5: EVOLUTIONARY ALGORITHMS Gabriela Ochoa http://www.cs.stir.ac.uk/~goc/ OUTLINE Optimisation problems Optimisation & search Two Examples The knapsack problem

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

International Journal of Digital Application & Contemporary research Website: (Volume 1, Issue 7, February 2013)

International Journal of Digital Application & Contemporary research Website:   (Volume 1, Issue 7, February 2013) Performance Analysis of GA and PSO over Economic Load Dispatch Problem Sakshi Rajpoot sakshirajpoot1988@gmail.com Dr. Sandeep Bhongade sandeepbhongade@rediffmail.com Abstract Economic Load dispatch problem

More information

Constrained Functions of N Variables: Non-Gradient Based Methods

Constrained Functions of N Variables: Non-Gradient Based Methods onstrained Functions of N Variables: Non-Gradient Based Methods Gerhard Venter Stellenbosch University Outline Outline onstrained Optimization Non-gradient based methods Genetic Algorithms (GA) Particle

More information

Proceedings of the First IEEE Conference on Evolutionary Computation - IEEE World Congress on Computational Intelligence, June

Proceedings of the First IEEE Conference on Evolutionary Computation - IEEE World Congress on Computational Intelligence, June Proceedings of the First IEEE Conference on Evolutionary Computation - IEEE World Congress on Computational Intelligence, June 26-July 2, 1994, Orlando, Florida, pp. 829-833. Dynamic Scheduling of Computer

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

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

Multi-objective pattern and feature selection by a genetic algorithm

Multi-objective pattern and feature selection by a genetic algorithm H. Ishibuchi, T. Nakashima: Multi-objective pattern and feature selection by a genetic algorithm, Proc. of Genetic and Evolutionary Computation Conference (Las Vegas, Nevada, U.S.A.) pp.1069-1076 (July

More information

Using Genetic Algorithms to Design Experiments: A Review

Using Genetic Algorithms to Design Experiments: A Review Using Genetic Algorithms to Design Experiments: A Review C. Devon Lin Department of Mathematics and Statistics, Queen s University Joint work with Christine M. Anderson-Cook, Michael S. Hamada, Lisa M.

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

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

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

Chapter 14 Global Search Algorithms

Chapter 14 Global Search Algorithms Chapter 14 Global Search Algorithms An Introduction to Optimization Spring, 2015 Wei-Ta Chu 1 Introduction We discuss various search methods that attempts to search throughout the entire feasible set.

More information

1 Lab + Hwk 5: Particle Swarm Optimization

1 Lab + Hwk 5: Particle Swarm Optimization 1 Lab + Hwk 5: Particle Swarm Optimization This laboratory requires the following equipment: C programming tools (gcc, make), already installed in GR B001 Webots simulation software Webots User Guide Webots

More information

Hybridization EVOLUTIONARY COMPUTING. Reasons for Hybridization - 1. Naming. Reasons for Hybridization - 3. Reasons for Hybridization - 2

Hybridization EVOLUTIONARY COMPUTING. Reasons for Hybridization - 1. Naming. Reasons for Hybridization - 3. Reasons for Hybridization - 2 Hybridization EVOLUTIONARY COMPUTING Hybrid Evolutionary Algorithms hybridization of an EA with local search techniques (commonly called memetic algorithms) EA+LS=MA constructive heuristics exact methods

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

A Real Coded Genetic Algorithm for Data Partitioning and Scheduling in Networks with Arbitrary Processor Release Time

A Real Coded Genetic Algorithm for Data Partitioning and Scheduling in Networks with Arbitrary Processor Release Time A Real Coded Genetic Algorithm for Data Partitioning and Scheduling in Networks with Arbitrary Processor Release Time S. Suresh 1, V. Mani 1, S. N. Omkar 1, and H. J. Kim 2 1 Department of Aerospace Engineering,

More information

5. Computational Geometry, Benchmarks and Algorithms for Rectangular and Irregular Packing. 6. Meta-heuristic Algorithms and Rectangular Packing

5. Computational Geometry, Benchmarks and Algorithms for Rectangular and Irregular Packing. 6. Meta-heuristic Algorithms and Rectangular Packing 1. Introduction 2. Cutting and Packing Problems 3. Optimisation Techniques 4. Automated Packing Techniques 5. Computational Geometry, Benchmarks and Algorithms for Rectangular and Irregular Packing 6.

More information

CHAPTER 6 ORTHOGONAL PARTICLE SWARM OPTIMIZATION

CHAPTER 6 ORTHOGONAL PARTICLE SWARM OPTIMIZATION 131 CHAPTER 6 ORTHOGONAL PARTICLE SWARM OPTIMIZATION 6.1 INTRODUCTION The Orthogonal arrays are helpful in guiding the heuristic algorithms to obtain a good solution when applied to NP-hard problems. This

More information

Graphical Approach to Solve the Transcendental Equations Salim Akhtar 1 Ms. Manisha Dawra 2

Graphical Approach to Solve the Transcendental Equations Salim Akhtar 1 Ms. Manisha Dawra 2 Graphical Approach to Solve the Transcendental Equations Salim Akhtar 1 Ms. Manisha Dawra 2 1 M.Tech. Scholar 2 Assistant Professor 1,2 Department of Computer Science & Engineering, 1,2 Al-Falah School

More information

Selection of Optimal Path in Routing Using Genetic Algorithm

Selection of Optimal Path in Routing Using Genetic Algorithm Selection of Optimal Path in Routing Using Genetic Algorithm Sachin Kumar Department of Computer Science and Applications CH. Devi Lal University, Sirsa, Haryana Avninder Singh Department of Computer Science

More information

Genetic Algorithms. Kang Zheng Karl Schober

Genetic Algorithms. Kang Zheng Karl Schober Genetic Algorithms Kang Zheng Karl Schober Genetic algorithm What is Genetic algorithm? A genetic algorithm (or GA) is a search technique used in computing to find true or approximate solutions to optimization

More information

Gen := 0. Create Initial Random Population. Termination Criterion Satisfied? Yes. Evaluate fitness of each individual in population.

Gen := 0. Create Initial Random Population. Termination Criterion Satisfied? Yes. Evaluate fitness of each individual in population. An Experimental Comparison of Genetic Programming and Inductive Logic Programming on Learning Recursive List Functions Lappoon R. Tang Mary Elaine Cali Raymond J. Mooney Department of Computer Sciences

More information

Outline. CS 6776 Evolutionary Computation. Numerical Optimization. Fitness Function. ,x 2. ) = x 2 1. , x , 5.0 x 1.

Outline. CS 6776 Evolutionary Computation. Numerical Optimization. Fitness Function. ,x 2. ) = x 2 1. , x , 5.0 x 1. Outline CS 6776 Evolutionary Computation January 21, 2014 Problem modeling includes representation design and Fitness Function definition. Fitness function: Unconstrained optimization/modeling Constrained

More information

Experimental Study on Bound Handling Techniques for Multi-Objective Particle Swarm Optimization

Experimental Study on Bound Handling Techniques for Multi-Objective Particle Swarm Optimization Experimental Study on Bound Handling Techniques for Multi-Objective Particle Swarm Optimization adfa, p. 1, 2011. Springer-Verlag Berlin Heidelberg 2011 Devang Agarwal and Deepak Sharma Department of Mechanical

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