Software Engineering Supervision Optimization Based On Multi Objective Particle Swarm Optimization Algorithm

In twenty-first Century, the software industry has been developed rapidly. It can be said that the size and complexity of software projects have been raised to a considerable extent. For such a large and rather complicated project, forming a good software engineering supervision system is of great significance. The multi-objective network plan of software engineering supervision mainly aims to guide the process of project supervision and to achieve the effective control over the project’s resources, construction period, cost and quality.At present, the research on software engineering supervision network aiming to solve optimization problem can hardly be found. While looking into the existing algorithms researches, on the one hand, the optimization effect is poor and solutions in the terms of diversity and convergence still need to be improved, on the other hand, only one or two markers in engineering have been studied. Considering this, this thesis has improved the multi-objective particle swarm optimization algorithm and established the multi-objective optimization model of software engineering supervision network planning in order to apply the improved algorithm to the optimization process of the network plan. The main contents include:(1) In view of the problems that convergence and distribution of solutions proposed by the software engineering supervision multi-objective network planning optimization method still need further improvement, this thesis improves the multi-objective particle swarm optimization algorithm by optimizing four main operators of the algorithm:the construction method of Pareto optimal solution set, the selection method of the global optimum value, the individual file update method and the selection method of individual extremum. And then, the improved multi-objective particle swarm optimization algorithm is tested by three standard multi-objective test functions:ZDT1, ZDT3, ZDT6. Compared with the genetic algorithm and the multi-objective particle swarm optimization algorithm based on crowding distance, it has been able to prove the convergence and distribution of the algorithm.(2) Aiming to solve the problem that only one or two goals of quality, cost, time and resources in the traditional software engineering supervision are optimized leading to inconsistency with the actual project, this thesis establishes a comprehensive mathematical optimization model and carries out a comprehensive optimization of multiple objectives based on the quality, cost, time and resources.(3) After applying the improved multi-objective particle swarm algorithm to the optimization process in the supervision of software engineering network planning, the algorithm process about supervision of software engineering network planning optimization has also been designed and the algorithm is applied to specific examples finally. All these tests has proved that this algorithm can get a relatively satisfactory solution.