Improvement And Application Of Gravitational Search Algorithm

In last few years, many swarm intelligence algorithms inspired by naturephenomenon have been proposed. These show by numerous experiments that thesealgorithms are good tool to solve complex single and multiple object optimizationproblems. All traditional optimization algorithms can not provide suitable preceptwhen they solve the problem of the optimization in high-dimensional space, sostudying the new optimization algorithm still is of its important value. Gravitationalsearch algorithm is new optimization algorithm for the same purpose. Due to itssimple principle and its high efficiency in solving some nonlinear functions in recentyears, it has become a research hot spot, and has been applied in some fields. Searchof the algorithm has two main aspects: One is how to improve its search accuracy, theother one is how to improve the convergence speed. The article is from the twoaspects to improve Gravitational search algorithm. Finally, the improvement ofgravitational search algorithm is applied in robots path planning.Gravitational search algorithm (GSA) based on Newton law of gravity and massinteractions is proposed and a new optimization algorithm. This paper is mainly aboutthe law of gravitation formula in the corresponding transformation to improve thegravitational search algorithm for improving the search accurate and accelerate speedof convergence.First, the principles and procedures of the GSA are summarized in the paper. Andin terms of defects in the standard GSA, such as resultant force deviation. This paperintroduces the concept of affinity, and affinity became the correct factor, improves theresultant force by introducing the affinity. So that it can weaken resultant deviation,and make particles moving in the expected direction, and make particles find optimalsolution, and then make search accurate at same time, accelerating convergence.Second, this paper inset the “black hole” operation to improved algorithm searchso that it not only improve accelerate speed of convergence but also search accurate.This article did some simulation experiments, and the results show that the twomethods have the effect.Finally, the gravitational search algorithm is applied in robot path planning, itincludes global path planning and Multi-robot path planning. Meanwhile, someinvalid particles will be change into random valid particles once again. So expandyour search range and keep part from get into local optimal. The path planning for mobile robots based PBGSA includes two steps: The first step is to establish afree-space mobile robot model, the second step is adopting PBGSA to fing out theglobal path. We do some theoretical simulations to verify the effectiveness of thealgorithm.