Jorge Carrizo, Fernando Tinetti and Pablo Moscato. CeTAD C.C. 75 ARGENTINA ABSTRACT

Size: px
Start display at page:

Download "Jorge Carrizo, Fernando Tinetti and Pablo Moscato. CeTAD C.C. 75 ARGENTINA ABSTRACT"

Transcription

1 A Computational Ecology for the Quadratic Assignment Problem Jorge Carrizo, Fernando Tinetti and Pablo Moscato CeTAD Universidad Nacional de La Plata C.C , La Plata ARGENTINA ABSTRACT A new heuristic method for the Quadratic Assignment Problem based on the use of a \Computational Ecology" and a modied version of Hillier-Connors heuristic HC-63 are presented in this paper. The method is intended to include another metaheuristic called Tabu Search to increase the diversity of solutions before recombining them to create new congurations which are used to restart the search. Its intrinsic parallelism was checked when implemented on a parallel computer (Transputer system).

2 Introduction The quadratic assignment problem (QAP) is a combinatorial optimization problem which is a representative of the important class of problems known as NP-hard, thus it is important to develop and test heuristics for it since membership in this class means that solution times are likely to be at least exponential with respect to N [16]. Heuristics are often considered by many scientists as \rules-of-thumb" which aid to the solution of specic problems but lack rigorousness. However, in combinatorial optimization where the concept of eectiveness [4] plays a fundamental role, heuristics should be regarded as part of an interplay with algorithms, an earlier stage in scientic discovery. The paper is organized as follows: the next subsections give a brief introduction to the maths involved in the formulation of the problem and some applications for it, section 2 describes the main ideas of the Steepest Descent Algorithm we have implemented and its relation with the HC-63, section 3 concerns with the population approach proposed, in section 4 results are shown, and nally section 5 presents our conclusions. The QAP is dened as follows: Denition and applications Given a vector of objects O = (o 1 ; o 2 ; : : : ; o N ) and a vector of data points (or places) Q = (q 1 ; q 2 ; : : : ; q N ) and a distance measure d : QQ! R +, the objective function we must minimize is?() given by?() = i=1 j=1 C ij D (i)(j) (1) where is a permutation of the set of indices (1; 2; : : : ; N) into the vector Q, (i) indicates the i th element of (the index of the point in Q assigned to object o i ) and C ij is a hypothesized similarity between objects o i and o j. The QAP is generally represented by two N N matrices C and D. The matrix C is called the \structure" or \cost" matrix and is dened by the terms C ij. The matrix D is called the \data" or \distance" matrix where each element D ij 2 D denotes the distance between points q i and q j. Thus D (i)(j) = d(q (i) ; q (j) ) (2) The QAP was rst formulated by Koopmans and Beckmann [9] followed by Gilmore [5] and Lawler [10] and since then it has found applications from data analysis to task scheduling on parallel computers. Applications also include the minimization of total wire length in electronic assemblies, location of machines, departments or oces within a plant so as to minimize transportation eorts and costs, ordering interrelated data on a disk, scheduling theory, distribution of medical services in a large hospital center, placement of logical modules in a chip, etc. In addition to this list, the Traveling Salesman Problem (TSP) (and with it all its applications) can be considered as a special case of the QAP in which the structure matrix is a cyclic permutation matrix.

3 Task Scheduling and Computational Load Balancing on Parallel Computers One of the advantages of studying heuristics for combinatorial optimization problems using large QAP instances is that many results can be evaluated using them as \test-beds" and the insights gained can be applied into other settings. This is the case for the load balancing issues arising in parallel computation. It can be viewed as a combinatorial optimization problem in which a number M of objects (Processes) must be assigned to a number N of places (Processors). With more generality we can consider it as a graph partitioning problem: given a graph G composed of 1. P vertices, from 0 to P? 1, which represent subproblems (tasks, processes, instructions, etc.) which will be processed concurrently, each one with an individual processing time (weight) w(p); 2. P (P? 1) communication links C(p; p 0 ) such that C(p; p 0 ) is the data amount (measured in bytes, words, etc.) to be sent from p to p 0 before the execution of the p 0 -th process. We wish to seek a partition of G in N subgraphs G 0 ; G 1 ; : : : G N?1 such that G = G 0 [ G 1 [ : : : [ G N?1 (3) where N is the quantity of nodes numbered from 0 up to N? 1 of a concurrent computer where the processes will be executed. 3. The computational load is well balanced, that is W n = X p w(p) Const (p 2 G n ); (4) 4. Communication time is minimal Comm = X p;p 0 C(p; p 0 ) t[n(p); n(p 0 )] = min (p 6= p 0 ) (5) where n(p) is the node that contains subprocess p and t[n; n 0 ] is the minimal time required to send data from the node n to node n 0. As we can see, the QAP is related with the communication part of the objective function. In this case, we will assume that communication time between processors only depends upon the distance between them so we can replace time by distance. We will mainly restrict ourselves to the study of QAP instances in which the P vertices lay on a grid (a rectangular array) due to the fact that we have large instances with known global minima. At rst glance, if no restrictions are given, the number of assignments of processes to processors is N! thus even for small sized problems (more than N = 15 or N = 20) the time needed for nding an optimal solution is prohibitive. Branch and bound was also applied but its implementation is also limited due to the computer time it requires [5] [10]. Iterative Improvement Strategies Heuristics may be classied in two main groups: 1.- Constructive. 2.- Iterative Improvement.

4 Constructive techniques start from either a partial or null assignation and they go on assigning new elements as long as restrictions or the objetive function constraints are satised. Iterative Improvement techniques start from an initial solution that includes all the elements to be placed and continue to make changes until a certain condition is reached. Among them we can remark those that belong to the SDPI class (Steepest Descent Pairwise Interchange), that is to take pairs of objects and try to interchange them regarding the cost function. For the QAP we must cite the H63 due to Hillier, the HC-66 due to Hillier and Connors [15], the CRAFT due to Armour and Bua and the `Biased Sampling' proposed by Nugent et al. With the exemption of the latter, they are all deterministic procedures. H63 and HC63-66 The HC63-66 is an iterative improvement heuristic (a modied version of the H63), which uses a so called \Move Desirability Table" (MDT) to select which pair of objects will be moved. The elements of the MDT indicate the cost changes that would result from \unilaterally" moving an object to a dierent location. This means that the net cost change of interchanging two elements is not evaluated at the time the MDT is made. The HC63-66 selects which objects will be moved just by inspection of the MDT which somewhat measures the individual \desire" of pair of objects to move. Once the objective function is evaluated, the move is accepted if it eectively produces a cost reduction (when the interchange considered is between objects at maximum distance, a minimum allowable cost reduction is imposed; a kind of \threshold"). The main dierence between the H63 and the HC63-66, is that the latter starts trying to move an object to locations which are as far as possible from the current location on the grid, while the former only considers up and down, and right or left moves. Eventually it also considers diagonal moves [15]. We can see other dierence in that H63 accepts \every" positive cost reduction, while H63-66 does not, some changes are discarded due to the \threshold". These heuristics are not only eective regarding the low-cost solutions it yields, but it also gains from the use of the MDT. In the case of H63, O(8 N) interchanges are tried instead of a naive O(N 2 ). The HC63-66 searches the O(N 2 ) neighborhood in an \ordered way" so most of the times this will conduce to a cost reduction interchange in less than O(N 2 ) attempts. We must recognize that this is an appealing characteristic but there is an obvious tradeo, since when we are near to local minima most of the moves that try to interchange the positions of objects far apart signicantly increase the cost. HC-91 When selecting what kind of method will be used for the rst stage, where local optimality is the goal, an iterative improvement procedure was chosen due to its solution quality and computational complexity. Among the collection of techniques described above, HC63-66 from Hillier & Connors was selected, showing a good performance in the average nal cost in several dierent instances of the QAP with sizes from ve to thirty department [15] along with a reduced computing time. This heuristic was modied for a new approach in which objects are selected according to its relative contribution to the objective function. Now we introduce a new heuristic which has many similarities with the previously cited methods. Up to a certain extent we can consider them as formal similarities. To avoid misinterpretations we will explicitly describe the main dierences. We will call the method \HC?91" since we think of it as a direct descendant of the strategies employed by Hillier & Connors. We have chosen a criterion to create the MDT which is a list based on the individual contribution of each object to the objective function. The MDT will give the interchange order between objects. For each object J.

5 MDT (J) = max J 0 (C JJ 0 D (J)(J 0 )) (J 6= J 0 ) (6) Another dierence can be found in the way the elements are selected to be interchanged. An auxiliary array is used to keep the sorting of the objects regarding the MDT. Using this auxiliary array, we take the rst object of the MDT, that is the object that has the biggest entry, and we try to interchage it with all the other elements. We follow the arbitrary order previously assigned, so if object o s has been selected, we try to interchange it with o 1 ; o 2 ; : : : until we nd an interchange that decreases the cost. We note o s as s and o v as v. Let's suppose we select object o v, then the decrement is given by (o s ; o v ) = +?? C ks D (k)(v) + C sk D (v)(k) + C ks D (k)(s)? C sk D (s)(k)? C kv D (k)(s) C vk D (s)(k) C kv D (k)(v) C vk D (v)(k) (7) From this formula we can see that the cost update requires O(N) operations. Finally, we do not forbid changes if they are smaller than a given threshold as implemented by Hillier & Connors. Each change that decreases the cost is performed no matter its magnitude. This lead us back from the 8 N of H63 to O(N 2 ) neighborhood, but in return it oers better solutions than the original method. We must remark that this approach also diers from CRAFT since not all pairwise exchanges are considered. This is true as long as an improvement can be made. In addition, we should note that the calculation of the cost reduction is performed in O(N) steps. Some possible changes in the denition of the MDT have been considered, with and without a non-zero minimal cost reduction step, but although they seemed promising the nal congurations found were not better than those found with the HC? 91. Tabu Search for the QAP Our basic heuristic strategy (HC-91) does not include backtracking, that is it halts after nding a local minimum. We have decided to include Tabu Search (TS) to explore neighboring solutions [6]. From a given conguration, TS seeks among the neighboring congurations (this depends on the move set), trying to nd that one that has the best evaluation (regarding the objective function, for example). If there are no improving moves possible, we are in a kind of local optimum and the one that least degrades the objective function is chosen. Two other elements are part of this \metaheuristic", the \tabu list" and the \aspiration criteria". The former is a data structure that forbids certain moves and attempts to avoid \cycling" while the second are rules that override the \tabu status" of some moves if they are relevant or interesting. An example can be those that lead to a better solution than the best found so far. We have not yet dened what are the elements of the Tabu list. Skorin-Kapov has selected them to be the exchange of two units that were exchanged during the preceding s iterations. Taillard has a dierent approach: \for every unit and location, the latest iteration at which the unit occupied that location is recorded. A move is tabu if it assigns both interchanged units to

6 locations that had occupied within the s most recent iterations." Nothing has been said about how long an element must remain \tabu" or the length of that list. Both issues have been studied in previous works [17] [18]. The implementation of a circular list of objects without a xed length can be found in Ref. [18]. Some of the instances studied in this paper have been also employed to study the behaviour of a \deterministic update" for simulated annealing [12]. The \Memetic" Algorithm With the name of \Genetic Algorithms" (GA), we recognize a number of methods, generally applied to operations research and articial intelligence problems, which make use of a \population approach" for the search. They borrowed the name from the analogies with biological mechanisms and they are part of what constitutes the eld of \evolutionary programming" [7]. Very recently, a number of techniques started to depart from traditional GA methods although they also shared two of the basic ingredients of GA, Reproduction and Crossover [14] [1]. These techniques have been tentatively called \memetic algorithms" due to the introduction of the word \meme" by R. Dawkins to designate the \unit of imitation". The \meme" would play the analogous role of the gene but in cultural evolution [11] [3]. These new approaches for optimization can also be viewed as \computational ecologies" where processes engage in competitive and cooperative behaviour [8] [13]. The memetic approach has proved to be a really powerful tool in the search of suboptimal solutions in other NP-hard problems like the TSP where interaction between individuals enhances the performance of individual heuristics [13]. Our basic strategy can be summarized with the following piece of pseudocode. Begin RANDOM INITIALIZATION; (* N a agents are created *) GenNum := 0; MaxGenNum := 40; Repeat HC-91 TO LOCAL MINIMA; (* all agents in the population *) SORT AGENTS; (* regarding the objective function *) TEST LOSS OF DIVERSITY; (* some agents may be discarded *) GENERATE NEW POPULATION; (* a new set of N a agents *) GenNum := GenNum + 1; Until (DiversityCrisisF lag = 1) or (GenNum > MaxGenNum) End. A certain number of agents (usually small, from 4 to 16 have been used in the tests) are initialized randomly and they undergo a period of local optimization until all of them reach local minima congurations. Then the agents are sorted regarding the value of the objective function. This ranking is used in generating the new population. Since we want to avoid \premature convergence" we test for diversity loss in each generation. We discard those agents which appear twice or more times in the population. That is, we eliminate exact duplicates but we leave two congurations if at least two objects have been assigned to dierent places. If the rst half of the sorted population has exactly the same conguration, then the population is judged to have a \diversity crisis" and we set DiversityCrisisF lag := 1. Selection criteria and Crossover The recombination operator (Crossover) is intended to help to explore the region of conguration space between good quality congurations. It is generally viewed as a mechanism that

7 interchanges information. In a \memetic" approach, the parents are well-developed congurations or even local optima, thus we expect it to exploit the correlation of local optima of the objective function landscape. Selection criteria for the parents are always needed. We also follow the guidelines of GA in selecting as parents the best agents (regarding the objective function) currently in the population. We took as parents, the best N a =2 dierent congurations. A new set composed of N a? 1 new agents is created at each generation using the crossover mechanism and the best local minima found in the previous step completes the list of N a agents. In this implementation, we have always selected the best local minima found in the previous step as one of the parents in order to have a fast convergence and to study general performance characteristics of the method but other type of selection criteria may be possible which would yield to better nal solutions. The process is continued until there is no diversity within the individuals. The best and the second best agents create three new agents, then the best with the third create two agents, then the best with the fourth creates two agents up to the eight agent in the list. For the QAP, H. Muhlenbein has recently introduced a recombination called N p -sexual voting recombination [14]. For each place, the object placements of N p parents are considered. If the same object is assigned to a particular place more times than a given threshold, then it is also assigned in the same place in the ospring. The remaining places will be completed with \mutations", that is random assignments of the other objects not yet placed. They presented results with N p = 7; 4 neighbours, the global best solution with a weight of two and the current conguration. We used the method proposed by Brown, Huntley & Spillane which is similar to the PMX operator used by Goldberg and Lingle [1]. It creates a new cyclic permutation O (for ospring) using two parent congurations (P 1 ; P 2 ) and the following procedure 1. Each of the parent structures is partitioned in three sections by randomly drawing two breaking points. 2. The second section of the O's structured is copied from the corresponding section of parent P The rst and third sections of the O structure are obtained by copying terms from corresponding positions in P 2 's structure which have not yet been assigned before. 4. The remaining empty positions in the rst and third sections of O are then obtained from a permutation of the terms not yet allocated. Results We have used two dierent population sizes in many runs of the heuristic. The population size is a parameter that takes an important role by giving better average results regarding the nal value of the cost function. The optimal size for the population (if any) was not studied in this paper. Instead, we select this value ad hoc regarding some preliminary runs. Although the SDPI heuristic of each individual is used many times during a run, we must remark that the population approach is dierent from a typical stochastic \multistart" algorithm. We have observed that the crossover operator contributes to drive the population towards better regions of conguration space. It is also remarkable that, as a consequence of that property, the time involved in using an individual SDPI decreases, that is the local minima are reached faster from recombined congurations in each generation. Seeking a comparison, we implemented a stochastic \multistart" algorithm using HC-91 a number of times which was approximately 5500; and the results were:

8 1. Average: Standard Deviation: Minimum: 6124 (The global minimum). More information about these results and the ones found adding Tabu Search can be found in [2]. With the population strategy, we made a preliminary study by doing 78 runs with a population size of 8, and the results found were: 1. Average: Standard Deviation: Minimum: 6124 The values given above are the best agent found at the end of a run, that is when a \diversity crisis" is reached. We should remark here that a run with a population size of 8 (or generally speaking, with a population size of N a ), is more expensive (in computer time) than 8 (or N a ) local minima found with the HC-91. The number of runs (5500 with HC-91 against 80 with the population approach), in some way \equalizes" the dierence regarding the computer time employed. As a deterministic SDPI method, the HC-91 has no freedom, given a starting point, it will end up in a certain, but a priori unknown, minimum. A characterization of its performance is possible given an idea of what will be the nal average cost. The bigger number of times it starts form a dierent point, the bigger the chances of nding a better solution. With this belief, more independent searches were added in order to improve the nal costs. So we try with a number of 16 HC-91 independent searches to embrace a wider range in the conguration space. As expected, better results were found reaching the known minima very easily and both the average cost and the standard deviation of the conguration cost have shown a reduction: 1. Average: Standard Deviation: Minimum: 6124 The selection of results was made in the same way that for the population size of 8. For a more accurate comparison between the results obtained with the two population sizes (8 and 16), we include frequency histograms made in the same way (the same cost values range: and the same \step" of 20) (see Fig. 1). At this time we should say that the increment in the population size from 8 to 16 will not double the computer time when the population is implemented on a sequential computer, (e.g. 2N a will not take twice the time taken by N a individual searches). We nd this type of approach to be particularly appropriate for implementation on parallel computers, like a Transputer network. To see the eect produced by discarding those agents which appear twice or more times in the population, we made some runs with a population size of 30 without such a \control". The population size of 30 seems to be better than the previously cited 8 and 16 (30 involves a wider range in the conguration space) and is appropriate to avoid \premature convergence". The results found were: 1. Average: 6152

9 2. Standard Deviation: Minimum: 6128 These results show the importance taken by such a control, because they are similar to those obtained with a population size of 8 in which the \duplicated" agents were discarded (and with less computer eort). Conclusions and further work As general conclusions on the results we can say: 1. The memetic approach searches better than a typical stochastic \multistart" approach. 2. The population size has an important eect on the results. 3. The control of diversity (and its relation with convergence) in the context of a memetic approach is necessary to obtain better results with a small population. Regarding the computer eort necessary to implement this kind of approach, we have mentioned before the relevance of a parallel implementation. There are (at least) three things to note about this: 1. SDPI heuristics are polynomial-time from the point of view of computational complexity. 2. If there are enough resources available, it will be possible to use a bigger population size, with at least two consequences: (a) A wider range in the conguration space will be explored. (b) A higher speed to reach good local minima will be obtained. 3. By improving the asynchronousness of the strategy we will improve the speedup of the global optimization process. These statements should be carefully studied to see the magnitude of each one, and their interaction. Early tests performed on problems of bigger sizes show an acceptable behavior of this heuristic. The Skorin-Kapov's instance of size 90 was tested and some results have been obtained. They were approximately 0:2 % above the best minimum found. This approach gives new breath to SDPI algorithms in the context of Memetic Algorithms. We can use dierent SDPI procedures in dierent agents and combine them (better said, the congurations they obtain) via the crossover operator to see how are the congurations in quality (cost) and how quickly they are found. These dierent SDPI procedures should be studied regarding their own behavior and how they interact in the population approach. The method described in this paper halts a run when a diversity crisis is reached. We are considering a method in which the population is restarted after the agents undergo a period of independent Tabu Search. This will increase diversity while preserving the congurations of the best local minima previously found. We expect that this blend would combine the strengths of both approaches. The inuence of population size must also be studied and it could be interesting to nd a relation between the population and the problem size. Acknowledgements We thank the rm MicroSistemas S.A. for the computer facilities and the CeTAD Director, A.A. Quijano for the support of this project. We specially thank J. Skorin-Kapov and E. Taillard. This work was conducted within the framework of the CICPBA's Program for Technology Transfer ( ).

10 REFERENCES [1] D. Brown, C.L. Huntley and A. Spillane, \A Parallel Genetic Heuristic for the Quadratic Assignment Problem", Proceedings of the Third International Conference on Genetic Algorithms, Fairfax, VA, (1989). [2] J. Carrizo, F.G. Tinetti, P. Moscato, \Heuristics for Computational Load Balancing and the Quadratic Assignment Problem", (in Spanish), Proceedings of the 4th Meeting in Information Processing and Control (RPIC '91, Buenos Aires, Argentina, 1991), pp , (Nov. 1991). [3] R. Dawkins, \The Selsh Gene", Oxford University Press, Oxford, (1976). [4] J. Edmonds, \Paths, Trees and Flowers", Can. J. Math., 17, pp , (1965). [5] P. Gilmore, \Optimal and Suboptimal Algorithms for the Quadratic Assignment Problem", J. of SIAM 10, pp , (1962). [6] F. Glover, \Tabu Search - Part I", ORSA Journal on Computing, 1 (3), pp , (1989). [7] J.H. Holland, \Adaptation in Natural and Articial Systems", University of Michigan Press, Ann Arbor, (1975). [8] B.A. Huberman ed. \The Ecology of Computation", (North-Holland, Amsterdam) (1988). [9] T. Koopmans and M. Beckmann, \Assignment Problems and the Location of Economic Activities", Econometrica 25, pp , (1957). [10] E. Lawler, \The Quadratic Assingnment Problem", Management Science 9, pp , (1963). [11] P. Moscato, \On Evolution, Search, Optimization, Genetic Algorithms and Martial Arts: Towards Memetic Algorithms", Caltech Concurrent Computation Program, Report C 3 P? 826, (1989). [12] P. Moscato and J.F. Fontanari, \Stochastic versus Deterministic Update in Simulated Annealing", Physics Letters A, 146, Number 4, pp , (21 May 1990). [13] P. Moscato and M. Norman, \A Competitive-Cooperative Approach to Complex Combinatorial Search", Caltech Concurrent Computation Program Report C 3 P? 790. Expanded version published in the Proceedings of the 20th JAIIO Conference, Buenos Aires, Argentina (August 1991). [14] H. Muhlenbein, \Parallel Genetic Algorithms, Population Genetics and Combinatorial Optimization", Proceedings of the Third International Conference on Genetic Algorithms, Fairfax, VA, (1989). [15] C.E. Nugent, T.E. Vollmann and J. Ruml, \An Experimental Comparison of Techniques for the Assignment of Facilities to Locations", Operations Research 16, pp , (1968) [16] S. Sahni and T. Gonzalez, \P-complete Approximation Problems", J. Assoc. Comput. Mach. 23, pp , (1976).

11 [17] Skorin-Kapov J., \Tabu Search Applied to the Quadratic Assignment Problem", ORSA Journal of Computing, Vol. 2, No. 1, pp , [18] Taillard E., \Robust Taboo Search for the Quadratic Assignment Problem", report ORWP 90/10, DMA, Swiss Federal Institute of Technology of Lausanne, Switzerland, 1990.

Genetic Hybrids for the. Quadratic Assignment Problem CHARLES FLEURENT AND JACQUES A. FERLAND. May 16, 1993

Genetic Hybrids for the. Quadratic Assignment Problem CHARLES FLEURENT AND JACQUES A. FERLAND. May 16, 1993 DIMACS Series in Discrete Mathematics and Theoretical Computer Science Volume 00, 0000 Genetic Hybrids for the Quadratic Assignment Problem CHARLES FLEURENT AND JACQUES A. FERLAND May 16, 1993 Abstract.

More information

ACO and other (meta)heuristics for CO

ACO and other (meta)heuristics for CO ACO and other (meta)heuristics for CO 32 33 Outline Notes on combinatorial optimization and algorithmic complexity Construction and modification metaheuristics: two complementary ways of searching a solution

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

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

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

Sparse Matrices Reordering using Evolutionary Algorithms: A Seeded Approach

Sparse Matrices Reordering using Evolutionary Algorithms: A Seeded Approach 1 Sparse Matrices Reordering using Evolutionary Algorithms: A Seeded Approach David Greiner, Gustavo Montero, Gabriel Winter Institute of Intelligent Systems and Numerical Applications in Engineering (IUSIANI)

More information

O(1) Delta Component Computation Technique for the Quadratic Assignment Problem

O(1) Delta Component Computation Technique for the Quadratic Assignment Problem O(1) Delta Component Computation Technique for the Quadratic Assignment Problem Sergey Podolsky, Yuri Zorin National Technical University of Ukraine Kyiv Polytechnic Institute Faculty of Applied Mathematics

More information

Using Genetic Algorithms to optimize ACS-TSP

Using Genetic Algorithms to optimize ACS-TSP Using Genetic Algorithms to optimize ACS-TSP Marcin L. Pilat and Tony White School of Computer Science, Carleton University, 1125 Colonel By Drive, Ottawa, ON, K1S 5B6, Canada {mpilat,arpwhite}@scs.carleton.ca

More information

Parallel Computing in Combinatorial Optimization

Parallel Computing in Combinatorial Optimization Parallel Computing in Combinatorial Optimization Bernard Gendron Université de Montréal gendron@iro.umontreal.ca Course Outline Objective: provide an overview of the current research on the design of parallel

More information

Telecommunication and Informatics University of North Carolina, Technical University of Gdansk Charlotte, NC 28223, USA

Telecommunication and Informatics University of North Carolina, Technical University of Gdansk Charlotte, NC 28223, USA A Decoder-based Evolutionary Algorithm for Constrained Parameter Optimization Problems S lawomir Kozie l 1 and Zbigniew Michalewicz 2 1 Department of Electronics, 2 Department of Computer Science, Telecommunication

More information

Genetic Algorithms for Solving. Open Shop Scheduling Problems. Sami Khuri and Sowmya Rao Miryala. San Jose State University.

Genetic Algorithms for Solving. Open Shop Scheduling Problems. Sami Khuri and Sowmya Rao Miryala. San Jose State University. Genetic Algorithms for Solving Open Shop Scheduling Problems Sami Khuri and Sowmya Rao Miryala Department of Mathematics and Computer Science San Jose State University San Jose, California 95192, USA khuri@cs.sjsu.edu

More information

A Parallel Architecture for the Generalized Traveling Salesman Problem

A Parallel Architecture for the Generalized Traveling Salesman Problem A Parallel Architecture for the Generalized Traveling Salesman Problem Max Scharrenbroich AMSC 663 Project Proposal Advisor: Dr. Bruce L. Golden R. H. Smith School of Business 1 Background and Introduction

More information

A tabu search based memetic algorithm for the max-mean dispersion problem

A tabu search based memetic algorithm for the max-mean dispersion problem A tabu search based memetic algorithm for the max-mean dispersion problem Xiangjing Lai a and Jin-Kao Hao a,b, a LERIA, Université d'angers, 2 Bd Lavoisier, 49045 Angers, France b Institut Universitaire

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

Non-deterministic Search techniques. Emma Hart

Non-deterministic Search techniques. Emma Hart Non-deterministic Search techniques Emma Hart Why do local search? Many real problems are too hard to solve with exact (deterministic) techniques Modern, non-deterministic techniques offer ways of getting

More information

Job Shop Scheduling Problem (JSSP) Genetic Algorithms Critical Block and DG distance Neighbourhood Search

Job Shop Scheduling Problem (JSSP) Genetic Algorithms Critical Block and DG distance Neighbourhood Search A JOB-SHOP SCHEDULING PROBLEM (JSSP) USING GENETIC ALGORITHM (GA) Mahanim Omar, Adam Baharum, Yahya Abu Hasan School of Mathematical Sciences, Universiti Sains Malaysia 11800 Penang, Malaysia Tel: (+)

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

SUITABLE CONFIGURATION OF EVOLUTIONARY ALGORITHM AS BASIS FOR EFFICIENT PROCESS PLANNING TOOL

SUITABLE CONFIGURATION OF EVOLUTIONARY ALGORITHM AS BASIS FOR EFFICIENT PROCESS PLANNING TOOL DAAAM INTERNATIONAL SCIENTIFIC BOOK 2015 pp. 135-142 Chapter 12 SUITABLE CONFIGURATION OF EVOLUTIONARY ALGORITHM AS BASIS FOR EFFICIENT PROCESS PLANNING TOOL JANKOWSKI, T. Abstract: The paper presents

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

Department of. Computer Science. Remapping Subpartitions of. Hyperspace Using Iterative. Genetic Search. Keith Mathias and Darrell Whitley

Department of. Computer Science. Remapping Subpartitions of. Hyperspace Using Iterative. Genetic Search. Keith Mathias and Darrell Whitley Department of Computer Science Remapping Subpartitions of Hyperspace Using Iterative Genetic Search Keith Mathias and Darrell Whitley Technical Report CS-4-11 January 7, 14 Colorado State University Remapping

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

HEURISTICS FOR THE NETWORK DESIGN PROBLEM

HEURISTICS FOR THE NETWORK DESIGN PROBLEM HEURISTICS FOR THE NETWORK DESIGN PROBLEM G. E. Cantarella Dept. of Civil Engineering University of Salerno E-mail: g.cantarella@unisa.it G. Pavone, A. Vitetta Dept. of Computer Science, Mathematics, Electronics

More information

The study of comparisons of three crossover operators in genetic algorithm for solving single machine scheduling problem. Quan OuYang, Hongyun XU a*

The study of comparisons of three crossover operators in genetic algorithm for solving single machine scheduling problem. Quan OuYang, Hongyun XU a* International Conference on Manufacturing Science and Engineering (ICMSE 2015) The study of comparisons of three crossover operators in genetic algorithm for solving single machine scheduling problem Quan

More information

A New Rank Based Version of the. Ant System. - A Computational Study. Bernd Bullnheimer. Richard F. Hartl. Christine Strau. Working Paper No.

A New Rank Based Version of the. Ant System. - A Computational Study. Bernd Bullnheimer. Richard F. Hartl. Christine Strau. Working Paper No. A New Rank Based Version of the Ant System - A Computational Study Bernd Bullnheimer Richard F. Hartl Christine Strau Working Paper No. 1 April 1997 April 1997 SFB `Adaptive Information Systems and Modelling

More information

Dept. of Computer Science. The eld of time series analysis and forecasting methods has signicantly changed in the last

Dept. of Computer Science. The eld of time series analysis and forecasting methods has signicantly changed in the last Model Identication and Parameter Estimation of ARMA Models by Means of Evolutionary Algorithms Susanne Rolf Dept. of Statistics University of Dortmund Germany Joachim Sprave y Dept. of Computer Science

More information

Tabu search and genetic algorithms: a comparative study between pure and hybrid agents in an A-teams approach

Tabu search and genetic algorithms: a comparative study between pure and hybrid agents in an A-teams approach Tabu search and genetic algorithms: a comparative study between pure and hybrid agents in an A-teams approach Carlos A. S. Passos (CenPRA) carlos.passos@cenpra.gov.br Daniel M. Aquino (UNICAMP, PIBIC/CNPq)

More information

International Journal of Current Research and Modern Education (IJCRME) ISSN (Online): & Impact Factor: Special Issue, NCFTCCPS -

International Journal of Current Research and Modern Education (IJCRME) ISSN (Online): & Impact Factor: Special Issue, NCFTCCPS - TO SOLVE ECONOMIC DISPATCH PROBLEM USING SFLA P. Sowmya* & Dr. S. P. Umayal** * PG Scholar, Department Electrical and Electronics Engineering, Muthayammal Engineering College, Rasipuram, Tamilnadu ** Dean

More information

Abstract. This paper shows how parallelism has been integrated into

Abstract. This paper shows how parallelism has been integrated into Parallel Optimisation in the SCOOP Library Per Kristian Nilsen 1 and Nicolas Prcovic 2 1 SINTEF Applied Mathematics, Box 124 Blindern, 0314 Oslo, NORWAY E-mail: pkn@math.sintef.no 2 CERMICS-INRIA Sophia

More information

A memetic algorithm for symmetric traveling salesman problem

A memetic algorithm for symmetric traveling salesman problem ISSN 1750-9653, England, UK International Journal of Management Science and Engineering Management Vol. 3 (2008) No. 4, pp. 275-283 A memetic algorithm for symmetric traveling salesman problem Keivan Ghoseiri

More information

SPATIAL OPTIMIZATION METHODS

SPATIAL OPTIMIZATION METHODS DELMELLE E. (2010). SPATIAL OPTIMIZATION METHODS. IN: B. WHARF (ED). ENCYCLOPEDIA OF HUMAN GEOGRAPHY: 2657-2659. SPATIAL OPTIMIZATION METHODS Spatial optimization is concerned with maximizing or minimizing

More information

Attractor of Local Search Space in the Traveling Salesman Problem

Attractor of Local Search Space in the Traveling Salesman Problem Attractor of Local Search Space in the Traveling Salesman Problem WEIQI LI School of Management University of Michigan - Flint 303 East Kearsley Street, Flint, Michigan 48502 U. S. A. Abstract: - A local

More information

A Genetic Algorithm for Multiprocessor Task Scheduling

A Genetic Algorithm for Multiprocessor Task Scheduling A Genetic Algorithm for Multiprocessor Task Scheduling Tashniba Kaiser, Olawale Jegede, Ken Ferens, Douglas Buchanan Dept. of Electrical and Computer Engineering, University of Manitoba, Winnipeg, MB,

More information

2 The Service Provision Problem The formulation given here can also be found in Tomasgard et al. [6]. That paper also details the background of the mo

2 The Service Provision Problem The formulation given here can also be found in Tomasgard et al. [6]. That paper also details the background of the mo Two-Stage Service Provision by Branch and Bound Shane Dye Department ofmanagement University of Canterbury Christchurch, New Zealand s.dye@mang.canterbury.ac.nz Asgeir Tomasgard SINTEF, Trondheim, Norway

More information

Optimizing the Sailing Route for Fixed Groundfish Survey Stations

Optimizing the Sailing Route for Fixed Groundfish Survey Stations International Council for the Exploration of the Sea CM 1996/D:17 Optimizing the Sailing Route for Fixed Groundfish Survey Stations Magnus Thor Jonsson Thomas Philip Runarsson Björn Ævar Steinarsson Presented

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

Journal of Global Optimization, 10, 1{40 (1997) A Discrete Lagrangian-Based Global-Search. Method for Solving Satisability Problems *

Journal of Global Optimization, 10, 1{40 (1997) A Discrete Lagrangian-Based Global-Search. Method for Solving Satisability Problems * Journal of Global Optimization, 10, 1{40 (1997) c 1997 Kluwer Academic Publishers, Boston. Manufactured in The Netherlands. A Discrete Lagrangian-Based Global-Search Method for Solving Satisability Problems

More information

Solving the Rural Postman Problem by Memetic Algorithms

Solving the Rural Postman Problem by Memetic Algorithms MIC 2001-4th Metaheuristics International Conference 679 Solving the Rural Postman Problem by Memetic Algorithms Ana Maria Rodrigues José Soeiro Ferreira INESC PORTO Rua José Falcão 110, 4050-315 Porto,

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

In Proc of 4th Int'l Conf on Parallel Problem Solving from Nature New Crossover Methods for Sequencing Problems 1 Tolga Asveren and Paul Molito

In Proc of 4th Int'l Conf on Parallel Problem Solving from Nature New Crossover Methods for Sequencing Problems 1 Tolga Asveren and Paul Molito 0 NEW CROSSOVER METHODS FOR SEQUENCING PROBLEMS In Proc of 4th Int'l Conf on Parallel Problem Solving from Nature 1996 1 New Crossover Methods for Sequencing Problems 1 Tolga Asveren and Paul Molitor Abstract

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

BI-OBJECTIVE EVOLUTIONARY ALGORITHM FOR FLEXIBLE JOB-SHOP SCHEDULING PROBLEM. Minimizing Make Span and the Total Workload of Machines

BI-OBJECTIVE EVOLUTIONARY ALGORITHM FOR FLEXIBLE JOB-SHOP SCHEDULING PROBLEM. Minimizing Make Span and the Total Workload of Machines International Journal of Mathematics and Computer Applications Research (IJMCAR) ISSN 2249-6955 Vol. 2 Issue 4 Dec - 2012 25-32 TJPRC Pvt. Ltd., BI-OBJECTIVE EVOLUTIONARY ALGORITHM FOR FLEXIBLE JOB-SHOP

More information

A B. A: sigmoid B: EBA (x0=0.03) C: EBA (x0=0.05) U

A B. A: sigmoid B: EBA (x0=0.03) C: EBA (x0=0.05) U Extending the Power and Capacity of Constraint Satisfaction Networks nchuan Zeng and Tony R. Martinez Computer Science Department, Brigham Young University, Provo, Utah 8460 Email: zengx@axon.cs.byu.edu,

More information

Homework 2: Search and Optimization

Homework 2: Search and Optimization Scott Chow ROB 537: Learning Based Control October 16, 2017 Homework 2: Search and Optimization 1 Introduction The Traveling Salesman Problem is a well-explored problem that has been shown to be NP-Complete.

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 Steady-State Genetic Algorithm for Traveling Salesman Problem with Pickup and Delivery

A Steady-State Genetic Algorithm for Traveling Salesman Problem with Pickup and Delivery A Steady-State Genetic Algorithm for Traveling Salesman Problem with Pickup and Delivery Monika Sharma 1, Deepak Sharma 2 1 Research Scholar Department of Computer Science and Engineering, NNSS SGI Samalkha,

More information

Heuristic Optimisation

Heuristic Optimisation Heuristic Optimisation Revision Lecture 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 Optimisation University

More information

A LOCAL SEARCH GENETIC ALGORITHM FOR THE JOB SHOP SCHEDULING PROBLEM

A LOCAL SEARCH GENETIC ALGORITHM FOR THE JOB SHOP SCHEDULING PROBLEM A LOCAL SEARCH GENETIC ALGORITHM FOR THE JOB SHOP SCHEDULING PROBLEM Kebabla Mebarek, Mouss Leila Hayat and Mouss Nadia Laboratoire d'automatique et productique, Université Hadj Lakhdar -Batna kebabla@yahoo.fr,

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

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

Overview of Tabu Search

Overview of Tabu Search Overview of Tabu Search The word tabu (or taboo) comes from Tongan, a language of Polynesia, where it was used by the aborigines of Tonga island to indicate things that cannot be touched because they are

More information

GENERATION OF GREY PATTERNS USING AN IMPROVED GENETIC- EVOLUTIONARY ALGORITHM: SOME NEW RESULTS

GENERATION OF GREY PATTERNS USING AN IMPROVED GENETIC- EVOLUTIONARY ALGORITHM: SOME NEW RESULTS ISSN 139 14X INFORMATION TECHNOLOGY AND CONTROL, 011, Vol.40, No.4 GENERATION OF GREY PATTERNS USING AN IMPROVED GENETIC- EVOLUTIONARY ALGORITHM: SOME NEW RESULTS Alfonsas Miseviius Kaunas University of

More information

Solving the C sum Permutation Flowshop Scheduling Problem by Genetic Local Search

Solving the C sum Permutation Flowshop Scheduling Problem by Genetic Local Search ICEC 98 (1998 IEEE International Conference on Evolutionary Computation) pp.230 234 Solving the C sum Permutation Flowshop Scheduling Problem by Genetic Local Search Takeshi Yamada, NTT Communication Science

More information

Solving Traveling Salesman Problem Using Combinational Evolutionary Algorithm

Solving Traveling Salesman Problem Using Combinational Evolutionary Algorithm Solving Traveling Salesman Problem Using Combinational Evolutionary Algorithm Mohammad Reza Bonyadi\ S.Mostafa Rahimi Azghadi^ and Hamed Shah Hosseini^ 1 Department of Electrical & Computer Engineering,

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

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

Solving Traveling Salesman Problem Using Parallel Genetic. Algorithm and Simulated Annealing

Solving Traveling Salesman Problem Using Parallel Genetic. Algorithm and Simulated Annealing Solving Traveling Salesman Problem Using Parallel Genetic Algorithm and Simulated Annealing Fan Yang May 18, 2010 Abstract The traveling salesman problem (TSP) is to find a tour of a given number of cities

More information

A New Heuristic for the Quadratic Assignment Problem

A New Heuristic for the Quadratic Assignment Problem JOURNAL OF APPLIED MATHEMATICS AND DECISION SCIENCES, 6(3), 143 153 Copyright c 2002, Lawrence Erlbaum Associates, Inc. A New Heuristic for the Quadratic Assignment Problem ZVI DREZNER zdrezner@fullerton.edu

More information

Artificial Intelligence

Artificial Intelligence Artificial Intelligence Informed Search and Exploration Chapter 4 (4.3 4.6) Searching: So Far We ve discussed how to build goal-based and utility-based agents that search to solve problems We ve also presented

More information

Algorithms & Complexity

Algorithms & Complexity Algorithms & Complexity Nicolas Stroppa - nstroppa@computing.dcu.ie CA313@Dublin City University. 2006-2007. November 21, 2006 Classification of Algorithms O(1): Run time is independent of the size of

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

APPLICATION OF THE FUZZY MIN-MAX NEURAL NETWORK CLASSIFIER TO PROBLEMS WITH CONTINUOUS AND DISCRETE ATTRIBUTES

APPLICATION OF THE FUZZY MIN-MAX NEURAL NETWORK CLASSIFIER TO PROBLEMS WITH CONTINUOUS AND DISCRETE ATTRIBUTES APPLICATION OF THE FUZZY MIN-MAX NEURAL NETWORK CLASSIFIER TO PROBLEMS WITH CONTINUOUS AND DISCRETE ATTRIBUTES A. Likas, K. Blekas and A. Stafylopatis National Technical University of Athens Department

More information

Lecture 4. Convexity Robust cost functions Optimizing non-convex functions. 3B1B Optimization Michaelmas 2017 A. Zisserman

Lecture 4. Convexity Robust cost functions Optimizing non-convex functions. 3B1B Optimization Michaelmas 2017 A. Zisserman Lecture 4 3B1B Optimization Michaelmas 2017 A. Zisserman Convexity Robust cost functions Optimizing non-convex functions grid search branch and bound simulated annealing evolutionary optimization The Optimization

More information

Solving Travelling Salesman Problem and Mapping to Solve Robot Motion Planning through Genetic Algorithm Principle

Solving Travelling Salesman Problem and Mapping to Solve Robot Motion Planning through Genetic Algorithm Principle Indian Journal of Science and Technology, Vol 8(35), DOI: 10.17485/ijst/2015/v8i35/86809, December 2015 ISSN (Print) : 0974-6846 ISSN (Online) : 0974-5645 Solving Travelling Salesman Problem and Mapping

More information

Genetic Algorithm for Dynamic Capacitated Minimum Spanning Tree

Genetic Algorithm for Dynamic Capacitated Minimum Spanning Tree 28 Genetic Algorithm for Dynamic Capacitated Minimum Spanning Tree 1 Tanu Gupta, 2 Anil Kumar 1 Research Scholar, IFTM, University, Moradabad, India. 2 Sr. Lecturer, KIMT, Moradabad, India. Abstract Many

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

A Genetic Algorithm for Database Query Optimization. University of Wisconsin West Dayton Street. that satises

A Genetic Algorithm for Database Query Optimization. University of Wisconsin West Dayton Street. that satises A Genetic Algorithm for Database Query Optimization Kristin Bennett Michael C. Ferris Computer Sciences Department University of Wisconsin 1210 West Dayton Street Madison, Wisconsin 53706 Yannis E. Ioannidis

More information

ii

ii ACO Algorithms for the Quadratic Assignment Problem Thomas Stutzle 1 and Marco Dorigo IRIDIA Universite Libre de Bruxelles ftstutzle,mdorigog@ulb.ac.be 1 Currently on leave from FG Intellektik, TU Darmstadt.

More information

Genetic Algorithm for Dynamic Capacitated Minimum Spanning Tree

Genetic Algorithm for Dynamic Capacitated Minimum Spanning Tree Genetic Algorithm for Dynamic Capacitated Minimum Spanning Tree Rahul Mathur M.Tech (Purs.) BU, AJMER IMRAN KHAN Assistant Professor AIT, Ajmer VIKAS CHOUDHARY Assistant Professor AIT, Ajmer ABSTRACT:-Many

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

Evolutionary Algorithms and the Cardinality Constrained Portfolio Optimization Problem

Evolutionary Algorithms and the Cardinality Constrained Portfolio Optimization Problem Evolutionary Algorithms and the Cardinality Constrained Portfolio Optimization Problem Felix Streichert, Holger Ulmer, and Andreas Zell Center for Bioinformatics Tübingen (ZBIT), University of Tübingen,

More information

Simplicial Global Optimization

Simplicial Global Optimization Simplicial Global Optimization Julius Žilinskas Vilnius University, Lithuania September, 7 http://web.vu.lt/mii/j.zilinskas Global optimization Find f = min x A f (x) and x A, f (x ) = f, where A R n.

More information

Stochastic Approximation Algorithms for Number Partitioning

Stochastic Approximation Algorithms for Number Partitioning Stochastic Approximation Algorithms for Number Partitioning The Harvard community has made this article openly available. Please share how this access benefits you. Your story matters. Citation Accessed

More information

Fuzzy Inspired Hybrid Genetic Approach to Optimize Travelling Salesman Problem

Fuzzy Inspired Hybrid Genetic Approach to Optimize Travelling Salesman Problem Fuzzy Inspired Hybrid Genetic Approach to Optimize Travelling Salesman Problem Bindu Student, JMIT Radaur binduaahuja@gmail.com Mrs. Pinki Tanwar Asstt. Prof, CSE, JMIT Radaur pinki.tanwar@gmail.com Abstract

More information

Using Penalties instead of Rewards: Solving OCST Problems with Problem-Specific Guided Local Search

Using Penalties instead of Rewards: Solving OCST Problems with Problem-Specific Guided Local Search Using Penalties instead of Rewards: Solving OCST Problems with Problem-Specific Guided Local Search Wolfgang Steitz, Franz Rothlauf Working Paper 01/2011 March 2011 Working Papers in Information Systems

More information

6. Tabu Search 6.1 Basic Concepts. Fall 2010 Instructor: Dr. Masoud Yaghini

6. Tabu Search 6.1 Basic Concepts. Fall 2010 Instructor: Dr. Masoud Yaghini 6. Tabu Search 6.1 Basic Concepts Fall 2010 Instructor: Dr. Masoud Yaghini Outline Tabu Search: Part 1 Introduction Illustrative Problems Search Space Neighborhood Structure Tabus Aspiration Criteria Termination

More information

Modified Order Crossover (OX) Operator

Modified Order Crossover (OX) Operator Modified Order Crossover (OX) Operator Ms. Monica Sehrawat 1 N.C. College of Engineering, Israna Panipat, Haryana, INDIA. Mr. Sukhvir Singh 2 N.C. College of Engineering, Israna Panipat, Haryana, INDIA.

More information

Simple mechanisms for escaping from local optima:

Simple mechanisms for escaping from local optima: The methods we have seen so far are iterative improvement methods, that is, they get stuck in local optima. Simple mechanisms for escaping from local optima: I Restart: re-initialise search whenever a

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

Heuristic Optimization Introduction and Simple Heuristics

Heuristic Optimization Introduction and Simple Heuristics Heuristic Optimization Introduction and Simple Heuristics José M PEÑA (jmpena@fi.upm.es) (Universidad Politécnica de Madrid) 1 Outline 1. What are optimization problems? 2. Exhaustive vs. Heuristic approaches

More information

Preliminary Background Tabu Search Genetic Algorithm

Preliminary Background Tabu Search Genetic Algorithm Preliminary Background Tabu Search Genetic Algorithm Faculty of Information Technology University of Science Vietnam National University of Ho Chi Minh City March 2010 Problem used to illustrate General

More information

Automatic Generation of Test Case based on GATS Algorithm *

Automatic Generation of Test Case based on GATS Algorithm * Automatic Generation of Test Case based on GATS Algorithm * Xiajiong Shen and Qian Wang Institute of Data and Knowledge Engineering Henan University Kaifeng, Henan Province 475001, China shenxj@henu.edu.cn

More information

Effective probabilistic stopping rules for randomized metaheuristics: GRASP implementations

Effective probabilistic stopping rules for randomized metaheuristics: GRASP implementations Effective probabilistic stopping rules for randomized metaheuristics: GRASP implementations Celso C. Ribeiro Isabel Rosseti Reinaldo C. Souza Universidade Federal Fluminense, Brazil July 2012 1/45 Contents

More information

A Two-Dimensional Mapping for the Traveling Salesman Problem

A Two-Dimensional Mapping for the Traveling Salesman Problem Computers Math. Apphc. Vol. 26, No. 12, pp. 65-73, 1993 0898-1221/93 $6.00 + 0.00 Printed in Great Britain. All rights reserved Copyright 1993 Pergarnon Press Ltd A Two-Dimensional Mapping for the Traveling

More information

Using Simple Ancestry to Deter Inbreeding for Persistent Genetic Algorithm Search

Using Simple Ancestry to Deter Inbreeding for Persistent Genetic Algorithm Search Using Simple Ancestry to Deter Inbreeding for Persistent Genetic Algorithm Search Aditya Wibowo and Peter Jamieson Dept. of Electrical and Computer Engineering Miami University Abstract In this work, we

More information

Application of Improved Discrete Particle Swarm Optimization in Logistics Distribution Routing Problem

Application of Improved Discrete Particle Swarm Optimization in Logistics Distribution Routing Problem Available online at www.sciencedirect.com Procedia Engineering 15 (2011) 3673 3677 Advanced in Control Engineeringand Information Science Application of Improved Discrete Particle Swarm Optimization in

More information

Comparison of TSP Algorithms

Comparison of TSP Algorithms Comparison of TSP Algorithms Project for Models in Facilities Planning and Materials Handling December 1998 Participants: Byung-In Kim Jae-Ik Shim Min Zhang Executive Summary Our purpose in this term project

More information

2ND INTERNATIONAL CONFERENCE ON METAHEURISTICS - MIC97 1. Graduate School of Engineering, Kyoto University

2ND INTERNATIONAL CONFERENCE ON METAHEURISTICS - MIC97 1. Graduate School of Engineering, Kyoto University 2ND INTERNATIONAL CONFERENCE ON METAHEURISTICS - MIC97 1 A Variable Depth Search Algorithm for the Generalized Assignment Problem Mutsunori Yagiura 1, Takashi Yamaguchi 1 and Toshihide Ibaraki 1 1 Department

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

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

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

A Randomized Algorithm for Minimizing User Disturbance Due to Changes in Cellular Technology

A Randomized Algorithm for Minimizing User Disturbance Due to Changes in Cellular Technology A Randomized Algorithm for Minimizing User Disturbance Due to Changes in Cellular Technology Carlos A. S. OLIVEIRA CAO Lab, Dept. of ISE, University of Florida Gainesville, FL 32611, USA David PAOLINI

More information

CPSC 320 Sample Solution, Playing with Graphs!

CPSC 320 Sample Solution, Playing with Graphs! CPSC 320 Sample Solution, Playing with Graphs! September 23, 2017 Today we practice reasoning about graphs by playing with two new terms. These terms/concepts are useful in themselves but not tremendously

More information

An Ant Approach to the Flow Shop Problem

An Ant Approach to the Flow Shop Problem An Ant Approach to the Flow Shop Problem Thomas Stützle TU Darmstadt, Computer Science Department Alexanderstr. 10, 64283 Darmstadt Phone: +49-6151-166651, Fax +49-6151-165326 email: stuetzle@informatik.tu-darmstadt.de

More information

CS 331: Artificial Intelligence Local Search 1. Tough real-world problems

CS 331: Artificial Intelligence Local Search 1. Tough real-world problems CS 331: Artificial Intelligence Local Search 1 1 Tough real-world problems Suppose you had to solve VLSI layout problems (minimize distance between components, unused space, etc.) Or schedule airlines

More information

Richard E. Korf. June 27, Abstract. divide them into two subsets, so that the sum of the numbers in

Richard E. Korf. June 27, Abstract. divide them into two subsets, so that the sum of the numbers in A Complete Anytime Algorithm for Number Partitioning Richard E. Korf Computer Science Department University of California, Los Angeles Los Angeles, Ca. 90095 korf@cs.ucla.edu June 27, 1997 Abstract Given

More information

METAHEURISTICS. Introduction. Introduction. Nature of metaheuristics. Local improvement procedure. Example: objective function

METAHEURISTICS. Introduction. Introduction. Nature of metaheuristics. Local improvement procedure. Example: objective function Introduction METAHEURISTICS Some problems are so complicated that are not possible to solve for an optimal solution. In these problems, it is still important to find a good feasible solution close to the

More information

Introduction to Optimization

Introduction to Optimization Introduction to Optimization Approximation Algorithms and Heuristics November 6, 2015 École Centrale Paris, Châtenay-Malabry, France Dimo Brockhoff INRIA Lille Nord Europe 2 Exercise: The Knapsack Problem

More information

An Application of Genetic Algorithm for Auto-body Panel Die-design Case Library Based on Grid

An Application of Genetic Algorithm for Auto-body Panel Die-design Case Library Based on Grid An Application of Genetic Algorithm for Auto-body Panel Die-design Case Library Based on Grid Demin Wang 2, Hong Zhu 1, and Xin Liu 2 1 College of Computer Science and Technology, Jilin University, Changchun

More information

Enumeration of Full Graphs: Onset of the Asymptotic Region. Department of Mathematics. Massachusetts Institute of Technology. Cambridge, MA 02139

Enumeration of Full Graphs: Onset of the Asymptotic Region. Department of Mathematics. Massachusetts Institute of Technology. Cambridge, MA 02139 Enumeration of Full Graphs: Onset of the Asymptotic Region L. J. Cowen D. J. Kleitman y F. Lasaga D. E. Sussman Department of Mathematics Massachusetts Institute of Technology Cambridge, MA 02139 Abstract

More information

3. Genetic local search for Earth observation satellites operations scheduling

3. Genetic local search for Earth observation satellites operations scheduling Distance preserving recombination operator for Earth observation satellites operations scheduling Andrzej Jaszkiewicz Institute of Computing Science, Poznan University of Technology ul. Piotrowo 3a, 60-965

More information