Adaptive step mechanism in Glowworm Swarm Optimization

Glowworm Swarm Optimization Algorithm (GSO) is one of new swarm intelligence optimization algorithms in recent years. Its main idea comes from the cooperative behavior source among individuals during the process of courtship and foraging. In this paper, in order to improve convergence speed in the late iteration, avoid the algorithm falling into local optimum, and reduce isolated nodes, the Adaptive Step Mechanism Glowworm Swarm Optimization (ASMGSO) is proposed. The main idea of ASMGSO algorithm is as follows: (1) On the basis of SMGSO algorithm, isolated nodes carry out bunching operator first, that is to say they are moving to the central position of the group. If the new position is not better than the current position, then isolated nodes perform mutation operation. (2) At the same time, the fixed step mechanism has been improved. Each individual has its step size which is adapted to itself in the iterative process. The step size is not only affected by the node's neighborhood density but also affected by the distance to the moving target node. The effectiveness of the proposed ASMGSO algorithm is verified through several classic test functions and application in Distance Vector-HOP.