Multi-Criteria Decision Making - Pareto Front Optimization Strategy for Solving Multi-Objective Problems
Adarsh Kesireddy, Luis Rodolfo García Carrillo, José Baca · 2020
To date, Multi-Objective Optimization (MOO) has been viewed in terms of both exploration and exploitation of the solution search area. Accordingly, the main contribution of this paper relies on the exploitation of the solution space, rather than the exploration of the solutions set. Exploration of the solution sets is performed with a combination of the Pareto front and Evolutionary Algorithm (EA). For exploitation in the search environment, Multi-Criteria Decision Making (MCDM) is used to pick the best performing agent/policy from the group of agents/policies. The novelty behind the methodology proposed in this paper is the introduction of Decision Making during the evolution of the policies, named Multi-Criteria Decision Making - Pareto Front (M-PF) optimization. The performance of the proposed solution is compared under numerical simulations against popular techniques such as Non-Dominated Sorting Algorithm -II (NSGA-II) and Non-Dominated Sorting Algorithm-III (NSGA-III).