Java Implementation of a Parameter-less Evolutionary Portfolio
This work addresses the parameter tuning challenge for users of evolutionary algorithms, but it is incremental as it builds on existing parameter-less methods.
The authors tackled the problem of parameter tuning in evolutionary algorithms by implementing a portfolio of parameter-less evolutionary algorithms in Java, which adaptively selects algorithms based on runtime performance criteria. Initial experiments showed that this portfolio can solve various problem classes without prior parameter setting, with an acceptable increase in computational effort.
The Java implementation of a portfolio of parameter-less evolutionary algorithms is presented. The Parameter-less Evolutionary Portfolio implements a heuristic that performs adaptive selection of parameter-less evolutionary algorithms in accordance with performance criteria that are measured during running time. At present time, the portfolio includes three parameter-less evolutionary algorithms: Parameter-less Univariate Marginal Distribution Algorithm, Parameter-less Extended Compact Genetic Algorithm, and Parameter-less Hierarchical Bayesian Optimization Algorithm. Initial experiments showed that the parameter-less portfolio can solve various classes of problems without the need for any prior parameter setting technique and with an increase in computational effort that can be considered acceptable.