Facility Location Model Solving Based On AFSA Hybrid Algorithm

Facility location is a classic Problem which is of great significance for reality. Theexploration of solving facility location model effectively has been the concern of manyscholars. On basis of predecessors’ methods and experiences, this paper puts forward a facilitylocation model on random demand based on AFSA hybrid intelligence algorithm, which is thecombination of AFSA and BP neural network. The algorithm has good convergence, highlyefficiency and is easy to realize and it also provides a new thinking and method for the facilitylocation model.The main contents are as follows:(1)Firstly, this paper makes analysis on the actual significance of facility location interms of its history, status quo and necessity and introduces several heuristic arithmeticpossibilities to attain random models.Then it puts forward AFSA hybrid intelligence algorithmused in this paper.(2)Secondly, it introduces three models of facility location due to random demand,which includes the building and solving of the model and analyzes the three models from theperspective of Math. The result shows that the method not only has too much computational,but also it can not get the integral value accurately. Thus, we search another method namedrandom simulation on basis of Matlab.(3)The third part is about random simulation which based on the fundamental theoriesand definitions. It explores the steps for random simulation variables, random simulationexpectation, probability, optimism value and pessimism value and works out the input andoutput data for the cost function out of the random simulation expectation model, randomchance restriction model and casual relevant chance model.(4)Furthermore, the paper focuses on the approach of the cost function and introducesthe relevant definitions of features of BP neural network and then testifies the approach ofincertitude function with the output and input function via BP neural network.(5)The last part analyzes the hybrid intelligence algorithm and makes improvement onthe traditional Artificial Fish arithmetic. Furthermore, data samples are employed to testify theastringency, based on which neural network and AFSA are combined to be a new hybridalgorithm to attain facility location model to verify its feasibility and practicability. The resultshows that the algorithm has good convergence, highly efficiency and is easy to realize.