Fish School Search Algorithm for Constrained Optimization
This work addresses constrained optimization for researchers and practitioners, but it is incremental as it builds on existing Fish School Search methods with minor modifications.
The authors tackled constrained optimization problems by applying a niching swarm metaheuristic, specifically the wFSS algorithm, and proposed variations with different constraint handling procedures. Results showed the approach could handle heavily constrained problems and achieve results comparable to state-of-the-art algorithms, but noted difficulties in fitness convergence due to local search operators in certain geometrical scenarios.
In this work we investigate the effectiveness of the application of niching able swarm metaheuristic approaches in order to solve constrained optimization problems. Sub-swarms are used in order to allow the achievement of many feasible regions to be exploited in terms of fitness function. The niching approach employed was wFSS, a version of the Fish School Search algorithm devised specifically to deal with multi-modal search spaces. A base technique referred as wrFSS was conceived and three variations applying different constraint handling procedures were also proposed. Tests were performed in seven problems from CEC 2010 and a comparison with other approaches was carried out. Results show that the search strategy proposed is able to handle some heavily constrained problems and achieve results comparable to the state-of-the-art algorithms. However, we also observed that the local search operator present in wFSS and inherited by wrFSS makes the fitness convergence difficult when the feasible region presents some specific geometrical features.