Discrete Real-world Problems in a Black-box Optimization Benchmark
Sprache des Titels:
GECCO '18: Proceedings of the Genetic and Evolutionary Computation Conference Companion
Combinatorial optimization problems come in a wide variety of types but five common problem components can be identified. This categorization can aid the selection of interesting and diverse set of problems for inclusion in the combinatorial black-box problem benchmark. We suggest two real-world problems for inclusion into the benchmark. One is a transport-lot building problem and the other one is the clustered generalized quadratic assignment problem. We look into designing an interface for discrete black-box problems that can accommodate problems belonging to all of the described categories as well real-world problems that often feature multiple problem components. We describe three different interfaces for black-box problems, the first using a general encoding for all types of problems the second one using specialized encodings per problem type and the last one describes problems in terms of the available operators. We compare the strengths and weaknesses of the three designs.