This paper aims to compare two different parallel approaches (cooperative and competitive) of the SPEA2 for solving the environmental-economic dispatch problem. The idea is to solve the problem by executing the SPEA2 algorithm along with three different meta-heuristics (Genetic Algorithms, Particle Swarm Optimization, and Differential Evolution) to perform changes in the population. The different meta-heuristics work in parallel using two different approaches. The first one is the competitive approach, in which meta-heuristics compete for producing the best set of candidate solutions for solving the problem. Whereas, the cooperative approach selects the new population merging all individuals from all meta-heuristics, then selecting the solution set for the Pareto frontier. The proposal was implemented in C++ using MPI in a master-slave parallel model. Two study cases were used: the first one with six generators and the second one with forty generators. Results showed that the cooperative approach presented the best Pareto frontier for the case of 40 generators.
A Novel Parallel SPEA2 for Solving the Environmental-Economic Dispatch Problem: Competitive vs Cooperative Approach