6533b7d5fe1ef96bd1264f65

RESEARCH PRODUCT

Biased Modern Heuristics for the OCST Problem

Franz Rothlauf

subject

Mathematical optimizationFitness functionOperator (computer programming)Computer scienceSimulated annealingGenetic algorithmDesign elements and principlesRepresentation (mathematics)HeuristicsSpan tree

description

Biasing modern heuristics is an appropriate possibility in designing problem-specific and high-quality modern heuristics. If we have knowledge about a problem we can bias the design elements of modern heuristics, namely the representation and search operator, fitness function, the initial solution, or even the search strategy. This chapter presents a case study on how the performance of modern heuristics can be increased by biasing the design elements towards high-quality solutions. Results show that problem-specific and biased modern heuristics outperform standard variants and even for large problem instances high-quality solutions can be found.

yearjournalcountryeditionlanguage
2011-01-01
https://doi.org/10.1007/978-3-540-72962-4_8