A Memetic Algorithm Applied to the Optimal Design of a Planar Mechanism for Trajectory Tracking

Abstract

Memetic algorithms (MA), explored in recent literature, are hybrid metaheuristics formed by the synergistic combination of a population-based global search technique with one or more local search algorithms, which in turn can be exact or stochastic methods. Different versions of MAs have been developed, and although their use was focused originally on combinatorial optimization, nowadays there are memetic developments to solve a wide selection of numerical type problems: with or without constraints, mono or multi objective, static or dynamic, among others. This paper presents the design and application of a novel memetic algorithm, MemMABC, tested in a case study for optimizing the synthesis of a four-bar mechanism that follows a specific linear trajectory. The proposed method is based on the MABC algorithm as a global searcher, with the addition of a modified Random Walk as a local searcher. MABC is a modified version of the Artificial Bee Colony algorithm, adapted to handle design constraints by implementing the feasibility rules of Deb. Four-bar mechanisms are a good example of hard optimization problems, since they are used in a wide variety of industrial applications; simulation results show a high-precision control of the proposed trajectory for the designed mechanism, thus demonstrating that MemMABC can be applied successfully as a tool for solving real-world optimization cases.