Generation of Compliant Mechanisms using Hybrid Genetic Algorithm

Compliant mechanism is a single piece elastic structure which can deform to perform the assigned task. In this paper, the compliant mechanisms are evolved using a constraint based bi-objective optimization formulation which requires one user defined parameter (η). This user defined parameter limits a gap between a desired path and an actual path traced by the compliantmechanism. The non-linear and discrete optimization problem is solved using the hybrid genetic algorithm (GA) wherein domain specific initialization, two-dimensional crossover operator and repairing techniques are adopted. A bit-wise local search method is used with elitist non-dominated sorting genetic algorithm to further refine the compliant mechanisms. Parallel computations are performed on the master-slave architecture to reduce the computation time. A parametric study is carried out for η value which suggests a range to evolve topologically different compliant mechanisms. The applied and boundary conditions to the compliantmechanisms are considered variables that are evolved by the hybrid GA. The post-analysis of results unveil that the complaint mechanisms are always supported at unique location that can evolve the non-dominated solutions.