An Elevator Kinematics Optimization Algorithm based on a Large Neighborhood Search for Optimizing Simulated Industrial Problems

Abstract

The EKO-LNS algorithm is a hybrid of elevator kinematics optimization (EKO) and large neighborhood search (LNS) that allows the EKO algorithm to quickly escape local optima. To that end, the motion condition is improved by considering candidates that are even worse than the worst in memory but are still far enough away from local optima. The EKO-LNS first proposed is compared to selected metaheuristics algorithms on the task to solve standard industrial optimization problems from the literature. Three-bar truss, speed reducer, pressure vessel, and tension/compression spring design issues are among them. In terms of the quality of both optimal operating conditions and convergence behavior, numerical results show that the EKO-LNS metaheuristic algorithm outperforms or competes with the other metaheuristic algorithms.