Finite Markov chain analysis of genetic algorithms
Proceedings of the Second International Conference on Genetic Algorithms on Genetic algorithms and their application
Correlation length, isotropy and meta-stable states
Proceedings of the 16th annual international conference of the Center for Nonlinear Studies on Landscape paradigms in physics and biology : concepts, structures and dynamics: concepts, structures and dynamics
Ruggedness and neutrality—the NKp family of fitness landscapes
ALIFE Proceedings of the sixth international conference on Artificial life
Efficiency of Local Search with Multiple Local Optima
SIAM Journal on Discrete Mathematics
The Density of States - A Measure of the Difficulty of Optimisation Problems
PPSN IV Proceedings of the 4th International Conference on Parallel Problem Solving from Nature
Synthetic Neutrality for Artificial Evolution
AE '99 Selected Papers from the 4th European Conference on Artificial Evolution
Measuring the evolvability landscape to study neutrality
Proceedings of the 8th annual conference on Genetic and evolutionary computation
Mutation-crossover isomorphisms and the construction of discriminating functions
Evolutionary Computation
Algebraic theory of recombination spaces
Evolutionary Computation
A study of NK landscapes' basins and local optima networks
Proceedings of the 10th annual conference on Genetic and evolutionary computation
Evolutionary algorithm characterization in real parameter optimization problems
Applied Soft Computing
A behavior-based analysis of modal problems
Proceedings of the 15th annual conference companion on Genetic and evolutionary computation
ParadisEO-MO: from fitness landscape analysis to efficient local search algorithms
Journal of Heuristics
Hi-index | 0.00 |
The performances of evolutionary algorithms (genetics algorithms, genetic programming, etc.) or local search algotihms (Simulated annealing, tabu search, etc.) depends on the properties of seach space structure. One concept to analyse the search space is the fitness landscapes in which the problem to optimize and the search algorithm are taken into account. The fitness landscape is a graph where the nodes are the potential solutions. The study of the fitness landscape consists in analysing this graph or a relevant partition of this graph according to the dynamic or search difficulty. This tutorial will give an overview, after an historical review of concept of fitness landscape, of the different ways to define fitness landscape in the field of evolutionary computation. Following, the two mains geometries (multimodal and neutral landscapes) corresponding to two different partitions of the graph, meets in optimization problems and the dynamics of metaheuristics on these will be given. The relationship between problems difficulty and fitness landscapes measures (autocorrelation, FDC, neutral degree, etc.) or the properties of the local optima networks, studied in recent work, will be deeply analysed. Finally, the tutorial will conclude with a brief survey of open questions and the recent researchs on fitness landscapes.