Multipopulational Metaheuristic Approaches to Real-Parameter Optimization

摘要: Multipopulational metaheuristic methods have been used to solve a variety of problems. The use multiple populations evolved in parallel and exchanging data according particular communication strategy is known mitigate premature convergence, enlarge diversity the populations, generally improve results obtained by maintaining sole panmictic population candidate solutions. Moreover, multipopulational algorithms can be easily parallelized efficiently accelerated contemporary multicore distributed architectures. In this work, we study two populational real-parameter optimization metaheuristics traditional configuration, propose new heterogeneous approach. usefulness method briefly evaluated on experiments with several well test functions for optimization.