ABSTRACT
Search space characterisation is a field that strives to define properties of gradients with the general aim of finding the most suitable stochastic algorithms to solve the problems. Diagnostic Optimisation characterises the search landscape while the search progresses. In this work, we have improved Predictive Diagnostic Optimisation to reduce the cost of the local search by introducing a sampling procedure to explore the neighbourhood. The neigbhourhood is created by the swap operator and the sample size recorded during the search is shown to correlate with the known characteristics of the problems.
- R. E. Burkard, S. Karisch, and F. Rendl. Qaplib-a quadratic assignment problem library. European Journal of Operational Research, 55(1):115--119, 1991.Google ScholarCross Ref
- R. Marti, G. Reinelt, and A. Duarte. Optsicom project-library for linear ordering problem, 2009.Google Scholar
- I. Moser and M. Gheorghita. Combining search space diagnostics and optimisation. In Congress on Computational Intelligence, Proceedings, pages 897--904. IEEE, 2012.Google ScholarCross Ref
- T. Stützle. Local search algorithms for combinatorial problems - analysis, improvements, and new applications, volume 220 of DISKI. Infix, 1999.Google Scholar
- E. D. Taillard. Scheduling instances, 2005.Google Scholar
Index Terms
- Characterising fitness landscapes using predictive local search
Recommendations
On Stochastic Fitness Landscapes: Local Optimality and Fitness Landscape Analysis for Stochastic Search Operators
Parallel Problem Solving from Nature – PPSN XVIAbstractFitness landscape analysis is a well-established tool for gaining insights about optimization problems and informing about the behavior of local and evolutionary search algorithms. In the conventional definition of a fitness landscape, the ...
Using holey fitness landscapes to counteract premature convergence in evolutionary algorithms
GECCO '08: Proceedings of the 10th annual conference companion on Genetic and evolutionary computationPremature convergence is a persisting problem in evolutionary optimisation, in particular - genetic algorithms. While a number of methods exist to approach this issue, they usually require problem specific calibration or only partially resolve the issue,...
Fitness landscapes and graphs: multimodularity, ruggedness and neutrality
GECCO '13 Companion: Proceedings of the 15th annual conference companion on Genetic and evolutionary computationOne of the most commonly-used metaphors to describe the process of heuristic search methods in solving combinatorial optimization problems is that of a fitness landscape. The landscape metaphor appears most commonly in work related to evolutionary ...
Comments