Analysis On Policy In Software Maintenance Process And Research On Human Allocation Model

Software maintenance is assuming ever more a crucial role in the lifecycle of software due to the increase of software requirements and the high variability of software environment. With widespread use of software in work and life, people have higher requirement for the quality and efficiency of software maintenance. This thesis, taking Yalong River Hydropower Development Company information platform as background, conducts some research about planning software maintenance policy in software maintenance process and constructing a human allocation model of maintenance tasks.At first, this thesis begins with the priority determination of maintenance tasks which is the first step of the maintenance process. The priority judging principle is analyzed from different levels. According to the actual operation condition, the priority level can be obtained by analyzing the impact area and the urgency of the task. After then, the maintenance policies are researched about four types of maintenance namely corrective, adaptive, perfective and preventive maintenance. The software maintenance policy models are built to obtain the optimal maintenance time point of corrective maintenance and the best maintenance cycle of preventive maintenance. The maintenance time of adaptive maintenance and four kinds of maintenance strategy for perfective maintenance are analyzed by using qualitative analysis method. Then the maintenance task list is generated according to previous studies. A mathematical model of the tasks allocation and workers scheduling is constructed and an ACO based simulation algorithm is designed for it. The goal is completing all the tasks with shortest time and limited resources. At last, based on the previous research and the real operation of the Yalong River Hydropower Development Company information platform, this thesis lists some maintenance cases to show the optimal maintenance polices in the maintenance process. Moreover, simulation results of ACO verify that the model and the algorithm are effective.This thesis takes the real operation of software maintenance into consideration, not only makes analysis on the policy in each step of software maintenance process in details but also builds a human allocation model of maintenance tasks. All this research provides a reference for the software maintenance work in enterprises so that the maintenance tasks can be completed more efficiently and the cost can be reduced.