Collaborative Swarm Fruit Fly Optimization Algorithm And Its Application In Web Service Composition

Fruit Fly Optimization Algorithm (FOA) is a new swarm intelligence optimization algorithm which imitates the collaborative behavior of fruit fly when foraging.The basic idea of FOA is to calculate fruit fly taste concentration value according to the position of each fruit fly random flight and find the best taste concentration value. Then fruit flies find food with iterations and the optimization problem is solved. FOA has many advantages such as the calculation process is simply, easy to understand, and so on. However, FOA has its shortcomings. For example, FOA uses a fixed step which easily leads to local search ability and global search ability out of balance. Meanwhile, the choice of the initial position of FOA has a great influence on algorithm stability.To overcome these disadvantages and improve the algorithm performance, a Collaborative Swarm Fruit Fly Optimization Algorithm (CSFOA) is proposed. CSFOA aims to solve the problem of web service composition. The contribution of this work is summarized as follows:(1)To the disadvantages of FOA, a Collaborative Swarm Fruit Fly Optimization Algorithm (CSFOA) is proposed. Firstly, CSFOA adopt collaborative double subgroups and diminishing step, which effectively improves the optimization precision and convergence speed of the algorithm. Then CSFOA utilize searching coefficient h to control the selection of initial position so as to improve the convergence stability.(2)CSFOA is applied in the continuous function optimization problem and tested on 18 benchmark functions. This thesis compare CSFOA with other classic swarm intelligence optimization algorithms. The experimental results show that CSFOA has better globally searching capability, faster convergence speed, higher convergence precision and stability than FOA. Compared to IFOA, PSO and DE, especially in solving high dimensional functions, CSFOA has higher optimization accuracy and stability.(3)To verify the practical applied ability of CSFOA, this thesis apply CSFOA to solve discrete web service composition. According the location of the fruit flies, CSFOA finds each web service location in workflow, and utilizes fitness function to calculate the service composition quality. Compare the experimental results with FOA, PSO and DE, CSFOA has better solving precision and speed. Meanwhile, CSFOA is better than PSO and DE in stability.