Hybridizing Invasive Weed Optimization and Simulated Annealing Algorithm for High-dimensional Function Optimization

摘要

We herein propose an efficient algorithm (called IWOSA herein after) hybridizing invasive weed optimization (IWO for short) with the simulated annealing (SA) algorithm. The IWO is a new algorithm proposed to solve actual practical problems, which imitates the invasive behavior of weeds in nature. In the further research IWO algorithm did not show its efficiency in high-dimensional problems, and lacked directivity in the process of IWOSA iterations. To deal with this problem, we employed IWO to provide diversity to explore solution and Metropolis criterion of SA to provide more precise guidance, and tried to improve accuracy and convergence speed by these steps. To test the proposed algorithm, we compared IWOSA with original IWO through high-dimensional optimization benchmark functions. The computational results showed the efficiency of our algorithm.