Research On Dynamic Software Project Scheduling Based On Variable Neighborhood Discrete Particle Swarm Optimization

With the continuous development of the economy,the competition between the software industries is intensifying.Software companies need to develop new software to meet the growing needs of human work and life.It is extremely important to set up a reasonable software project scheduling program.In the face of software projects of ever-increasing scale,software project managers need to adopt intelligent means to provide efficient scheduling solutions in a short time,and comprehensively consider relevant factors to reduce the time and cost of software project development.As the most important resource in software project scheduling,employees' mobility will bring great uncertainty to software projects,making them more complicated than certain software project scheduling problems.This article focuses on the dynamic software project scheduling problem,considering the random departure of employees and the addition of new employees.A Markov chain is used to describe the process of random employee turnover,and Monte Carlo method is used to sample employee turnover.The addition of new employees,taking into account the addition of employees in normal and emergency situations,set up corresponding heuristic rules.This paper builds a dynamic software project scheduling model based on pre-scheduling and rescheduling strategies.In the pre-scheduling phase,all constraints are met,the employees are assigned corresponding tasks,and a pre-scheduling plan is generated.The re-scheduling phase takes into account the lack of skills caused by the employee's resignation,adds new employees to the software project scheduling process,and adjusts the scheduling plan for unfinished tasks.According to the form of software project scheduling scheme solution,the collective particle swarm algorithm is used to solve the SPSP problem,and the position,velocity and update formula of the particle swarm algorithm are redefined.In order to prevent the algorithm from falling into local optimum,and based on the characteristics of the problem,a variable neighborhood discrete particle swarm algorithm is proposed,and three neighborhood structures for software project scheduling are designed.The experimental results show that the proposed variable neighborhood discrete particle swarm optimization algorithm has higher efficiency and effectiveness,and can provide more intelligent decision-making for software project managers.