Convergence Analysis of Invasive Weed Optimization Algorithm and Its Application in Engineering

Abstract:

This paper introduces a novel numerical stochastic optimization algorithm,the invasive weed optimization (IWO),inspired from colonizing weeds,which mimics the robustness,adaptation and randomness of weeds in a simple but effective optimizing algorithm.Its global convergence is analyzed with Markov chain.Compared to other heuristic algorithms,the biggest advantage of IWO is its directed search based on the species of outstanding individuals within the group.Additionally,the offspring individuals are being randomly spread near their parents according to Gauss normal distribution with the standard deviation of the random function adjusted dynamically during the evolution process.Thus,the algorithm explores new areas aggressively to maintain the diversity of the species in the early and middle iterations,and then enhance the local search near optimal individuals in final iterations.Such mechanism ensures the steady convergence of the algorithm to global optimal solution.Simulation results of the optimal design of a typical complex machinery show that IWO algorithm can effectively search global optimum to avoid falling into a local optimal solution.