Differential Evolution Optimization Technique To Design
Gear Train System

Abstract

Determination of volume or center-to-center distance of a gear is an important issue in design of power transmission systems. The aim of this research work was to automate the design of gear drives by minimizing volume and center-to-center distance of gear trains. Differential Evolution Optimization technique was applied to a parallel axis gear train problems. Dynamic penalty to the objective function was also introduced for handling the constraints. Differential Evolution Optimization is metaheuristics search algorithm. It optimizes a problem by trying to improve a candidate solution iteratively. For this it takes a given measure of quality i.e. fitness fuction. DE does not require the problem to be differential as is usually required by classical optimization methods. It can be used on problems that are not even continuous. DE assumes a population of candidate solution and creates a new candidate solution by combining existing ones in accordance to its formula. It then compares the existing solution with the new candidate solution and keeps whichever has the best score. This process is repeated several times until the stopping criterion is met.