COMBINING EFO AND MVO: A PARALLEL HYBRID AND INFORMATION SHARING STRATEGY APPLIED TO OPTIMIZATION PROBLEMS

Abstract

This paper proposes a hybrid strategy combining the Electromagnetic Field Optimizer (EFO) and the Multi-Verse Optimizer (MVO) using a parallel execution approach with periodic information sharing, called EFO-MVO. This hybrid method leverages the advantages of each algorithm: EFO's strong local exploitation and MVO's effective global exploration. To evaluate the performance of EFO-MVO, we conducted experiments on six well-known benchmark functions: Sphere, Rastrigin, Rosenbrock, Ackley, Griewank, and Zakharov. The results demonstrate that EFO-MVO consistently achieves superior optimization outcomes compared to standalone EFO, MVO, and Genetic Algorithm (GA). The EFO-MVO method effectively avoids local optima and converges quickly to the global optimum, indicating its robust potential for complex optimization problems