Application Research Of Optimization Algorithm Based On Work-Flow

Optimization theory with its algorithm is an important branch of mathematics;it is concerned with the problem of which project is the best in many projects and how to find the best solution. At present, there are genetic algorithm, simulated annealing, ant colony algorithm, particle swarm algorithm, artificial neural network and so on. They are stochastic search iterative algorithms and no demand of optimization object. Because each algorithm has different mechanisms of search, different optimization characteristic and different application fields, they always depend on experience when selecting suitable algorithm for factual problems. In addition, the product of research in intelligence optimization is separate and lack of systematization at the present time. There appear lots of improvements of single algorithm and combination of multi-algorithms. This phenomenon is negative for developing this field.Therefore, different from previous methods that only select several functions to study performance of optimization algorithm, this paper will focus on the optimization character and algorithm flow, analysis the optimization efficiency and application fields when solving combination optimization problems, abstract algorithm flow with five key modules, initialization, calculation, estimation, update and output. After reasonable combination, we built the uniform flow structure of optimization algorithm;by analyzing the algorithm flow and work-flow, this paper studied the way of mapping the algorithm flow to work-flow, and modeling the algorithm flow based on work-flow theory;from the point of evaluate the structure performance of algorithm model, adopt Petri- net to analysis the model, which proved the unified flow model is soundness in structure;to evaluate the running performance of the algorithm model, build a combination of genetic algorithm and simulated annealing based on the unified model, adopt the time performance parameter and calculate performance parameter, the experiment proves that, the algorithm based on the unified model could reflect the advantage of original algorithm. The experiment proves conclusion of research: the unified flow model could reflect the character of algorithm in the best structure, support multi-algorithm combination.