AAFSA-RA:A Multi-Resource Allocation Method Based on an Advanced Artificial Fish Swarm Algorithm
Huan Ziqian · Xi'an Jiaotong Daxue xuebao · 2014
A muti-resource allocation method based on an advanced artificial fish swarm algorithm(AAFSA-RA)is proposed to solve the problem of multi-resource allocation among tasks in a resource-constrained environment.The muti-resource allocation is modeled,and is derived as a complex optimization problem with muti-constraints.Then an advanced artificial fish swarm algorithm(AAFSA)is proposed to improve the computational speed of the basic artificial fish swarm algorithm(AFSA).Random walks of artificial fishes are properly limited to improve prey behavior.A two-stage dynamic adjustment mechanism is introduced.The adjustment in the firststage is based on the evolutionary process of fish school,and the adjustment in the second-stage is based on the food concentration of single artificial fish and its changes.The life cycle of each artificial fish is controlled according to its food concentration.Finally,the problem of mutiresource allocation is solved using AAFSA.Simulation results show that AAFSA-RA steadily finds the global optimum value with a speed faster than the speed of existing algorithms based onAFSA;Comparisons with the traditional resource allocation methods show that the AAFSA-RA achieves higher precision without any special requirements on utility functions.