The Improvements And Applications Of The Glowworm Swarm Algortihm

The glowworm swarm optimization (GSO) algorithm is one of the intelligence algorithms, which is proposed by krishnanad and Ghose who come form India in 2005. The GSO algorithm is good at solving continuous optimization problem, which has been successfully used for collective robot, multi-modal function optimization, network sensor layout, cluster analysis and so on. Although The GSO algorithm is successfully applied in some areas, the algorithm has also many deficiencies through analyzing, such as time-consuming, the precision is not high and the extreme is possibly left and so on. It is needed to improve the performance from the basic theory, design of algorithm and so on. Besides, its applications need to be expanded the scope.This paper makes some researches of the GSO algorithm for the above problems, and obtains some good results, which are summarized as follows:(1) With the increase of the extreme points, the convergence speed and the computing accuracy of the glowworm swarm optimization (GSO) algorithm are low and not high. Aiming at the shortcomings of the GSO algorithm, this paper proposes a new improved algorithm of small-scale and multi-population glowworm swarm optimization (MPGSO).It is shown by simulation that, compared to GSO, the improved algorithm for solving multi-modal functions can not only obviously reduce the computing time, but also improve the computing accuracy.(2) In order to overcome the shortcomings, which glowworm swarm optimization (GSO) will miss the boundary extreme points or adjacent extreme points that are very close, this paper proposes two improvements. On the basis of the advantages of the GSO and the PSO, the glowworm - particle swarm hybrid optimization algorithm (GPSO) is designed. It shows that, through simulation results, GPSO has advantages as high precision, high speed, high stability, and will not fall into local optimum.(3) Aiming at the inefficiency of glowworm swarm optimization (GSO) algorithm for solving high-dimensional function problems, thus this paper presents a new improved glowworm swarm optimization algorithm with disturbance term (GSO-DT). It is shown by simulation that, the GSO-DT outperforms the SGSO than GSO in terms of efficiency, reliability, accuracy and stability. Especially, for the high-dimensional function optimization, the improved algorithm has clear superiority.(4) QoS-GSO algorithm is presented to solve the QoS constrained multicast routing problems. Our test shows that the algorithm can find optimal solution quickly and has better performance than GA and ACO. Furthermore, for the larger QoS constrained multicast routing problem,the algorithm can also quickly obtain the correct solution, which has good prospects of application.(5) The MOP-GSO algorithm is presented, which is for mutli-objective optimization.. It is shown by simulation that, MOP-GSO algorithm is effective to solve multi-objective optimization. Compared with NSGA2,it is better in the spread of the solutions.