MDE-ITMF and DEwI: Two New Multiple Solution Algorithms for Multimodal Optimization

TL;DR

This paper introduces two novel multipopulation algorithms, MDE-ITMF and DEwI, that effectively find multiple solutions in complex multimodal optimization problems using iterative penalization and differential evolution.

Contribution

The paper presents two new algorithms, MDE-ITMF and DEwI, combining iterative penalization and multipopulation strategies with differential evolution for comprehensive multimodal optimization.

Findings

Both methods successfully identified multiple solutions on benchmark functions.

MDE-ITMF and DEwI outperform some existing algorithms in solution diversity.

The approaches are promising tools for complex multimodal problems.

Abstract

Mathematical formulations of real world optimization studies frequently present characteristics such as non-linearity, discontinuity and high complexity. This class of problems may also exhibit a high number of global minimum/maximum points, especially for optimization problems arising from nonlinear algebraic systems (where null minima correspond to the solutions of the original algebraic system). Due to the multimodal nature of these functions, multipopulation methods have been employed in order to obtain the highest number of points of global minimum/maximum. In this work, two new approaches were analyzed, employing an iterative penalization technique and a multipopulation procedure---together with the Differential Evolution algorithm---devoted to obtain the full set of solutions for multimodal optimization problems. The first method proposed is the Multipopulation Differential Evolution with iterative technique of modification of the objective function, MDE-ITMF, and the second method proposed is the Differential Evolution with Initialization, DEwI. In this second proposal, the MDE-ITMF method is used as an initializer of the initial populations and from a given moment the Differential Evolution is used to solve the problem at hand. In both approaches, subpopulations evolve simultaneously throughout the iterative process. MDE-ITMF and DEwI methods were applied in a set of ten multimodal benchmark functions. Based on the results obtained, we can conclude that MDE-ITMF and DEwI are suitable and promising tools for multimodal optimization.