A Catfish Effect Inspired Harmony Search Algorithm for Optimization

摘要

This paper describes a Catfish Effect inspired harmony search algorithm. In the algorithm, two equal size harmony memories (HM) are used to cover all the variable range, and the new harmony is generated from each history one or from the whole possible solution set, following some probability rule. After the new harmony is obtained, the worst harmony in one HM will be evaluated by the new and the best harmony in the other HM, respectively. Which means that a better harmony has the effect that causing the weak HM to better themselves. These procedures proceed on until the nearoptimal solution vector is obtained. Numerical comparisons of the new algorithm with several existing harmony search algorithms are presented based on some standard benchmark optimization problems from CEC2005. The results show that the new algorithm is quite promising.