Formation de niches en optimisation continue multimodale L'optimisation continue multimodale a pour objectif de trouver tous les optima globaux d'une fonction. Dans ce contexte, plusieurs méthodes de niche (e.g. " crowding " ou " fitness sharing, 1989. ,
ontétéontété introduites pour la « formation et la préservation de sous-populations » stables [Mahfoud, 1995a; Mahfoud Chaque sous-population est concernée par un pic de la fonction multimodale et peutêtrepeutêtre interprétée intuitivement comme une sous-espèce qui exploite une " niché ecologique " . Par rapportàrapportà cette vision, en optimisation discrète il nous semble délicat d'encourager la populationàpopulationà se spécialiser sur des niches fixes, le processus de recherche devrait passer en permanence d'une régionrégionà une autre. Pour les fonctions multimodales, les niches correspondent aux quelques pics de la fonction objectif. De cette façon, le croisementà croisementà l'intérieur d'une seule sous-population est même encouragé, 1993. ,
Fitness sharing est une technique de niche très populaire basée sur le principe suivant : si deux individus sont distants d'au moins ? share (le rayon de niche, ou " sharing parameter " ), alors leurs valeurs d'adaptation sont mutuellement pénalisées, 1987. ,
Bien que cette approche soit différente de notre stratégie, un point commun est la définition d'un rayon de niche approprié ? similairèsimilairè a notré ecart minimum cible R. Si notre stratégie refuse l'insertion d'enfants sur un rayon de R autour de tout individu, le " fitness sharing " ne fait que pénaliser les valeurs de la fonction d'adaptation pour des individus distants de moins de R. Dans les deux cas, si un rayon inapproprié est utilisé, l'efficacité est considérablement diminuée. Nous avons trouvé une bonne valeur de rayon en utilisant une analyse du groupement des meilleurs individus ; en " fitness sharing, des théories sur le rayon sont disponibles (voir, 1989. ,
Mahfoud's deterministic crowding [Mahfoud, 1995a], voir aussi [Cedeño and Vemuri, 1999]) o` u l'on essaie de construire des sous-populations de niche « en forçant de nouveaux individusàindividusà remplacer ceux qui sont similaires génétiquement De cette manì ere, un nouvel individu remplace toujours un individu de sa propre sous-population et il minimise le changement sur la population existante, i.e. il préserve la diversité déjà existante. Cela peut rappeler notre remplacement direct. Mais, tandis que notre opération 1. No-Div (Colonnes 5 et 6) : Evo?Div sans stratégie de diversité ? i.e. tout enfant est accepté et l'individu le plus mauvais estéliminéàestéliminéestéliminéà l'´ etape de remplacement, No-Div est significativement plus mauvais. Bien qu'il puisse encore trouver quelques solutions, les taux de réussite stables d'Evo?Div sont perdus parce que la diversité n'existe plus, 1975. ,
Evo?Div avec un croisement de base, similaire avec GPX [Galinier and Hao, 1999] (voir Section 5.3.2) Bien que Basic-Cross puisse trouver 70% des solutions atteintes par Evo?Div, son plus grandprobì eme est qu'il rend l'algorithme plus lent. Même en ignorant les 30% d'´ echec, Basic-Cross exige environ dix fois plus de temps de calcul (voir DSJC1000.1) et c'est pour cette raison qu'il obtient des taux de réussite inférieurs, 1000. ,
algorithme détecte que la condition n'est pas satisfaite (il retourne impossible en O(|S|), voir Algorithme 6.1 annexéannexéà la page 111), et, ensuite ,
cet algorithme a permis d'améliorer considérablement la vitesse d'exécution (e.g. le nombre d'itérations par minute) des algorithmes TS?Div et TS?Int, Cela demanderait au moins O(|S| + k 2 ) opérations et, pour une instance comme, p.0, 1000. ,
impossible " après O(|S|) opérations. Dans cette situation, l'algorithme hongrois est ensuite appliqué, mais la complexité totale de calcul n'est pas alourdie par ces O(|S|) opérations, car O(|S|) compte peu dans une somme de complexités comme O(|S| + k 2 ) + O(|S|) = O(|S| + k 2 ) De plus, l'algorithme proposé peutêtrepeutêtre utile même si les conditions ne sont que partiellement satisfaites sur un sous ensemble de S (voir Section 6.4) Dans une telle situation ,
certaines idées de l'algorithme proposé peuventêtrepeuventêtre utiles pour résoudre desprobì emes généraux d'affectation, définis avec des matrices creuses ,
Algorithme 6.1 présente dans cette annexe un simple pseudocode ,
La fonction de remplacement (´ elimination) d'Evo ,
An application of iterated local search to graph coloring ,
Stochastic local search algorithms for the graph colouring problem, Handbook of approximation algorithms and metaheuristics, pp.63-64, 2007. ,
Stochastic Local Search Methods for Highly Constrained Combinatorial Optimisation Problems, 2005. ,
The priority-based coloring approach to register allocation, ACM Transactions on Programming Languages and Systems, vol.12, issue.4, pp.501-536, 1990. ,
DOI : 10.1145/88616.88621
An algorithm for the chromatic number of a graph, The Computer Journal, vol.14, issue.1, pp.38-39, 1971. ,
DOI : 10.1093/comjnl/14.1.38
Handling multiple objectives with particle swarm optimization, IEEE Transactions on Evolutionary Computation, vol.8, issue.3, 2004. ,
DOI : 10.1109/TEVC.2004.826067
Ants can colour graphs, Journal of the Operational Research Society, vol.48, issue.3, pp.295-305, 1997. ,
DOI : 10.1057/palgrave.jors.2600357
Embedding a sequential procedure within an evolutionary algorithm for coloring problems in graphs, Journal of Heuristics, vol.7, issue.5, pp.105-128, 1995. ,
DOI : 10.1007/BF02430368
Frozen development in graph coloring, Theoretical Computer Science, vol.265, issue.1-2, pp.227-264, 2001. ,
DOI : 10.1016/S0304-3975(01)00164-5
Exploring the k-colorable landscape with iterated greedy, Cliques, Coloring, and Satisfiability Second DIMACS Implementation Challenge [Johnson and Trick, pp.345-284, 1996. ,
Central partition for a partition-distance and strong pattern graphs, REVSTAT?Statistical Journal, vol.2, issue.2, pp.127-143, 2004. ,
Tabu Search for Graph Coloring, T-Colorings and Set T-Colorings, Meta-Heuristics Advances and Trends in Local Search Paradigms for Optimization, pp.77-92, 1998. ,
DOI : 10.1007/978-1-4615-5775-3_6
Méthodes heuristiques pour la coloration, la T-coloration et l'affectation de fréquences, Du and Pardalos, p.60, 1998. ,
Handbook of Combinatorial Optimization, p.76, 1998. ,
Genetic algorithms and grouping problems, pp.24-25, 1996. ,
Genetic and hybrid algorithms for graph coloring, Annals of Operations Research, vol.22, issue.3, pp.437-461, 1996. ,
DOI : 10.1007/BF02125407
Object-oriented implementation of heuristic search methods for graph coloring, maximum clique, and satisfiability, Cliques, Coloring, and Satisfiability Second DIMACS Implementation Challenge [Johnson and Trick, pp.619-652, 1996. ,
An Evolutionary Annealing Approach to Graph Coloring, Proceedings of EvoWorkshops2001?Applications of Evolutionary Computing, pp.120-129, 1992. ,
DOI : 10.1007/3-540-45365-2_13
Knowledge discovery in databases : An overview. AI Magazine, Freisleben and Merz, pp.57-70, 1992. ,
A genetic local search algorithm for solving symmetric and asymmetric traveling salesman problems, Proceedings of IEEE International Conference on Evolutionary Computation, pp.616-621, 1996. ,
DOI : 10.1109/ICEC.1996.542671
A minimal-state processing search algorithm for graph coloring problems IEICE Transactions on Fundamentals of Electronics, Galinier and Hao, pp.1420-1430, 1999. ,
Hybrid evolutionary algorithms for graph coloring, Galinier and Hertz, pp.379-397, 1999. ,
DOI : 10.1023/A:1009823419804
A survey of local search methods for graph coloring. Computers and operations research, pp.2547-2562, 2006. ,
An adaptive memory algorithm for the k-coloring problem, Discrete Applied Mathematics, vol.156, issue.2, pp.267-279, 1999. ,
DOI : 10.1016/j.dam.2006.07.017
Etude des métahueristiques pour la satisfaction de contraintes et la coloration, p.13, 1999. ,
A graph coloring model for a feasibility problem in monthly crew scheduling with preferential bidding, Gamst and Rave, pp.2384-2395, 1982. ,
DOI : 10.1016/j.cor.2005.09.010
On frequency assignment in mobile automatic telephone systems, Proceedings of the IEEE Global Communications Conference, pp.309-315, 1979. ,
Computers and Intractability : A Guide to the Theory of NP-completeness, Freeman, p.13, 1979. ,
What Color Is Your Jacobian? Graph Coloring for Computing Derivatives, SIAM Review, vol.47, issue.4, pp.629-671, 1998. ,
DOI : 10.1137/S0036144504444711
Local optima topology for the 3-sat problem, Cahiers du GERAD, vol.59, pp.98-68, 1998. ,
A polynomially searchable exponential neighbourhood for graph colouring, Journal of the Operational Research Society, vol.8, issue.3, pp.324-330, 2002. ,
DOI : 10.1007/BF02122551
Bag rationalisation for a food manufacturer, Journal of the Operational Research Society, vol.53, issue.5, pp.544-551, 1997. ,
DOI : 10.1057/palgrave.jors.2601316
Tabu Search, p.30, 1996. ,
URL : https://hal.archives-ouvertes.fr/hal-01389283
Coloring by tabu branch and bound, Cliques, Coloring, and Satisfiability Second DIMACS Implementation Challenge [Johnson and Trick, pp.285-307, 1986. ,
Future paths for integer programming and links to artificial intelligence. Computers and Operations research, pp.533-549, 1986. ,
Genetic algorithms with sharing for multimodal R ´ EFÉRENCESEF´EFÉRENCES BIBLIOGRAPHIQUES function optimization, Proceedings of the Second International Conference on Genetic Algorithms, pp.41-49, 1987. ,
Partition-distance: A problem and class of perfect graphs arising in clustering, Information Processing Letters, vol.82, issue.3, pp.159-164, 2002. ,
DOI : 10.1016/S0020-0190(01)00263-0
Frequency assignment: Theory and applications, Proceedings of the IEEE, pp.1497-1514, 1980. ,
DOI : 10.1109/PROC.1980.11899
Scatter Search for Graph Coloring, In Artificial Evolution LNCS, vol.2310, pp.168-179, 2001. ,
DOI : 10.1007/3-540-46033-0_14
An Analysis of Solution Properties of the Graph Coloring Problem, Metaheuristics computer decision-making, pp.325-345, 2004. ,
DOI : 10.1007/978-1-4757-4137-7_15
The steepest ascent mildest descent heuristic for combinatorial programming ,
Metaheuristiques pour l'optimisation combinatoire et l'affectation sous contraintes. Revue d'Intelligence Artificielle, pp.283-324, 1999. ,
Die Tabu-Methoden zur Graphenf??rbung, Computing, vol.21, issue.4, pp.345-351, 1987. ,
DOI : 10.1007/BF02239976
Local optima topology for the k-coloring problem, Discrete Applied Mathematics, vol.49, issue.1-3, pp.257-280, 1994. ,
DOI : 10.1016/0166-218X(94)90212-7
Variable space search for graph coloring, Discrete Applied Mathematics, vol.156, issue.13 ,
DOI : 10.1016/j.dam.2008.03.022
Stochastic Local Search Foundations & Applications, 2004. ,
Long path problems, PPSN 94, pp.149-158, 1994. ,
DOI : 10.1007/3-540-58484-6_259
Cliques, Coloring, and Satisfiability Second DIMACS Implementation Challenge, volume 26 of DIMACS series in Discrete Mathematics and Theoretical Computer Science, 1996. ,
Properties of Fitness Functions and Search Landscapes, Theoretical aspects of evolutionary computing, pp.175-206, 1972. ,
DOI : 10.1007/978-3-662-04448-3_8
Reducibility among combinatorial problems ,
List-graph colouring for multiple depot vehicle scheduling, International Journal of Mathematics in Operational Research, vol.1, issue.1/2, pp.228-245, 1979. ,
DOI : 10.1504/IJMOR.2009.022883
A graph coloring algorithm for large scheduling problems, Journal of Research of the National Bureau of Standards, vol.84, issue.6, pp.489-503, 1979. ,
DOI : 10.6028/jres.084.024
Experiments with parallel graph coloring heuristics and applications of graph coloring, Cliques, Coloring, and Satisfiability Second DI- MACS Implementation Challenge [Johnson and Trick, pp.309-334, 1996. ,
Robust Graph Coloring for Uncertain Supply Chain Management, Proceedings of the 38th Annual Hawaii International Conference on System Sciences ,
DOI : 10.1109/HICSS.2005.526
A critical element-guided perturbation strategy for iterated local search, EvoCOP, pp.1-12, 2009. ,
A memetic algorithm for graph coloring, European Journal of Operational Research, vol.203, issue.1, pp.241-250, 2006. ,
DOI : 10.1016/j.ejor.2009.07.016
An exact method for graph coloring, Computers & Operations Research, vol.33, issue.8, pp.2189-2207, 1994. ,
DOI : 10.1016/j.cor.2005.01.008
URL : https://hal.archives-ouvertes.fr/hal-00783637
A survey on vertex coloring problems, International Transactions in Operational Research, vol.39, issue.12, pp.92-98, 2008. ,
DOI : 10.1111/j.1475-3995.2009.00696.x
A Metaheuristic Approach for the Vertex Coloring Problem, INFORMS Journal on Computing, vol.20, issue.2, pp.302-82, 1999. ,
DOI : 10.1287/ijoc.1070.0245
Improving graph colouring with linear programming and genetic algorithms, Proceedings of Eurogen99, pp.113-118, 1999. ,
Contrôle générique de Paramètres pour les AlgorithmesÉvolutionnairesAlgorithmes´AlgorithmesÉvolutionnaires, Mehrotra and Trick, p.13, 1996. ,
A Column Generation Approach for Graph Coloring, Méndez-Díaz and Zabala, pp.344-354, 1996. ,
DOI : 10.1287/ijoc.8.4.344
A Branch-and-Cut algorithm for graph coloring, Merz and Freisleben, pp.826-847, 2006. ,
DOI : 10.1016/j.dam.2005.05.022
Fitness landscape analysis and memetic algorithms for the quadratic assignment problem, Merz and Freisleben, pp.337-352, 2000. ,
Fitness Landscapes, Memetic Algorithms, and Greedy Operators for Graph Bipartitioning, Evolutionary Computation, vol.5, issue.1, pp.61-91, 2000. ,
DOI : 10.1109/4235.585893
Advanced Fitness Landscape Analysis and the Performance of Memetic Algorithms, Evolutionary Computation, vol.13, issue.4, pp.303-325, 1996. ,
DOI : 10.1109/4235.887236
Genetic algorithms with dynamic niche sharing for multimodalfunction optimization, Proceedings of IEEE International Conference on Evolutionary Computation, pp.786-791, 1996. ,
Distributed coloration neighborhood search, Cliques, Coloring, and Satisfiability Second DIMACS Implementation Challenge [Johnson and Trick, pp.335-358, 1996. ,
An Experimental Investigation of Iterated Local Search for Coloring Graphs, EvoWorkshops, volume 2279 of LNCS, pp.121-130, 2002. ,
DOI : 10.1007/3-540-46004-7_13
Breaking instance-independent symmetries in exact graph coloring, Proceedings Design, Automation and Test in Europe Conference and Exhibition, pp.191-224, 2006. ,
DOI : 10.1109/DATE.2004.1268868
Genetic Algorithms, Path Relinking, and the Flowshop Sequencing Problem, Evolutionary Computation, vol.64, issue.1, pp.45-60, 1998. ,
DOI : 10.1016/0377-2217(89)90383-4
Feature Article???Genetic Algorithms for the Operations Researcher, INFORMS Journal on Computing, vol.9, issue.3, pp.231-250, 1997. ,
DOI : 10.1287/ijoc.9.3.231
Sur quelques aspects mathématiques desprobì emes de classification automatique, Mathématiques et Sciences Humaines, vol.82, issue.4, pp.20-175, 1965. ,
An effective two-stage simulated annealing algorithm for the minimum linear arrangement problem, Computers & Operations Research, vol.35, issue.10, pp.3331-3346, 2008. ,
DOI : 10.1016/j.cor.2007.03.001
An improved simulated annealing algorithm for bandwidth minimization, European Journal of Operational Research, vol.185, issue.3, pp.1319-1335, 2008. ,
DOI : 10.1016/j.ejor.2005.12.052
Artificial Intelligence : A Modern Approach, 2002. ,
The depth and width of local minima in discrete solution spaces, Discrete Applied Mathematics, vol.56, issue.1, pp.75-82, 1995. ,
DOI : 10.1016/0166-218X(93)E0134-K
An improved algorithm for exact graph coloring, Cliques, Coloring, and Satisfiability Second DIMACS Implementation Challenge [Johnson and Trick, pp.359-376, 1996. ,
Ants and Graph Coloring, proceedings of the International Conference Artificial Neural Nets and Genetic Algorithms in Prague, Czech Republic, pp.276-279, 2001. ,
DOI : 10.1007/978-3-7091-6230-9_68
Dcga a diversity control oriented genetic algorithm, Proceedings of the 9th IEEE International Conference on Tools with Artificial Intelligence, pp.367-374, 1997. ,
Searching for Diverse, Cooperative Populations with Genetic Algorithms, Evolutionary Computation, vol.2, issue.2, pp.127-149, 1993. ,
DOI : 10.1162/evco.1993.1.2.127
MAPM: memetic algorithms with population management, Computers & Operations Research, vol.33, issue.5, pp.1214-1225, 2006. ,
DOI : 10.1016/j.cor.2004.09.011
The landscape of the traveling salesman problem, Physics Letters A, vol.161, issue.4, pp.337-344, 1992. ,
DOI : 10.1016/0375-9601(92)90557-3
How the landscape of random job shop scheduling instances depends on the ratio of jobs to machines, Journal of Artificial Intelligence Research, vol.26, pp.247-287, 2006. ,
Distance based hybrid genetic algorithm: an application for the graph coloring problem, Proceedings of the 1999 Congress on Evolutionary Computation-CEC99 (Cat. No. 99TH8406), 1999. ,
DOI : 10.1109/CEC.1999.785564
Breaking the search space symmetry in partitioning problems, Theoretical Computer Science, vol.378, issue.1, pp.78-86, 2007. ,
DOI : 10.1016/j.tcs.2007.01.023
landscape case, Physical Review E, vol.78, issue.6, pp.66114-2007, 2008. ,
DOI : 10.1103/PhysRevE.78.066114
URL : https://hal.archives-ouvertes.fr/hal-00354804
A Large Neighborhood Search Heuristic for Graph Coloring, LNCS, vol.4510, issue.88, pp.346-360, 2002. ,
DOI : 10.1007/978-3-540-72397-4_25
Diversity-Guided Evolutionary Algorithms, PPSN VII, pp.462-471, 2002. ,
DOI : 10.1007/3-540-45712-7_45
Guided local search, Handbook of metaheuristics, pp.185-218, 2003. ,
Analyse et Résolution Approchées deProbì emes d'Optimisation Combinatoire : Application auProbì eme de Coloration de Graphe, 2004. ,
An upper bound for the chromatic number of a graph and its application to timetabling problems, The Computer Journal, vol.10, issue.1, pp.85-86, 1967. ,
DOI : 10.1093/comjnl/10.1.85
An Efficient Algorithm for Computing the Partition Distance ,
An Evolutionary Approach with Diversity Guarantee and Well-Informed Grouping Recombination for Graph Colo- ring ,
Position Guided Tabu Search for Graph Coloring, LNCS, vol.5851, pp.148-162, 2009. ,
A Study of Evaluation Functions for the Graph K-Coloring Problem, Selected papers from EA'07 (8th International Conference on Artificial Evolution Publications Conférences et rencontres nationaux ? Daniel Porumbel, pp.124-135, 2007. ,
DOI : 10.1007/978-3-540-79305-2_11
Un algorithme Guidé pour la Coloration de Graphe, ExposéExposéà l'atelier Journées Graphes et Algorithmes 09, 2009. ,
Recherche Tabou renforcée pour la coloration de graphe, ROADEF 2008 ? 9éme conférence de la ROADEF, 2008. ,
Reinforced Tabu Search for Graph Coloring, ExposéExposéà l'´ equipe d'optimisation, 2008. ,