Fuzzy Multi-objective Requirements for NRP Based on Particle Swarm Optimization

摘要

In software engineering, the development of software products raises a new set of development requirements each time. Considering the interaction between requirements, how to select an optimal subset of requirements becomes an important problem. In this paper, a fast method of requirements optimization is proposed, which can select an optimal subset from the next release of product development requirements under the limitation of user satisfactions and cost. The multiple requirements in this paper are limited by user satisfaction and cost. We mainly make the following contributions: (1) We define this problem as multi-objective problem for optimization. (2) Then particle swarm optimization (PSO) algorithm is used to adjust the convergence parameters of multiple object to search the optimal solution quickly. (3) Finally, the results of the algorithm are evaluated by using NDS number and time of multi-objective problem through fuzzy simulation data. Experimental results show that the algorithm is efficient and reliable, and can help developers make reasonable decisions.