Design and Optimization of Distribution Routes Using Evolutionary Strategy and Type-1 Singleton Neuro-Fuzzy Systems
Pascual Noradino Montes Dorantes, Pedro Henoc Ireta Sánchez, José Manuel Velarde Cantú, Ernesto Liñán García, Gerardo Maximiliano Méndez · IEEE Latin America Transactions · 2018
This paper presents a new proposal to design and to optimize the routes by hybrid intelligent model using evolutionary strategies and Adaptive Network Fuzzy Inference System (ANFIS). The hybridization uses an evolutionary computing model to create random a priori routes in the first step. The model serves as the inputs to the singleton type-1 ANFIS system that provides the knowledge and changes the connections between customers or sub-routes for distances by approximation. The generated routes solved by ANFIS model in the second phase will be optimized in the last phase of the proposal with the evolutionary strategy. The goal of this proposal is to provide a simple way to optimize routes without a hard mathematical model that produces an optimal or sub-optimal solution without an optimization function reducing the complexities of classical approaches.