Research On Task Assignment Of Software Maintenance Personnel Based On Learning Curve

Internet technology has developed vigorously in recent years.Internet users has increased significantly in the meanwhile.Software maintenance projects rise and increase day by day.The increase of software functions which leads to the increase of structural complexity,the lack of project documents,the loss of maintenance team personnel,the replacement of managers and the diversity of business modules and the complexity of business requirements increases the maintenance difficulty to software testers.Although the team members’ familiarity with the business requirements of each module in the maintenance project is uneven,the iteration time of the maintenance project is relatively abundant,so that team members can continue learning in the regression test process by task assignment,so as to improve their familiarity with unknown modules and improve the technical level at the same time.From the competency,combined with the industry,enterprise,stage of project particularity,combined with the ability evaluation indicators from employees in the software industry and team in company,ability evaluation indicators of software testers will be remained first.The indicators with high importance and high frequency in the project maintenance stage will be filterred out by the questionnaire for testers,which are determined as final ability evalution indicators.Then the project managers determine the weight of each evaluation indicator.Final score weight is get by AHP,and the personal ability evaluation model establishes.From the business module side,the regression testing of each module with the characteristics of manual operation and repeatability has a learning effect.By collecting the historical data of multiple personnel regression tests,using the cumulative average method,using SPSS software to help calculate the learning rate of each module.The number of executions of each module and the time it takes determine the learning curve function.Learning curve is used to predict the time it takes when perform the regression testing for this module next time.The personnel ability is extracted from the evaluation results of personal ability evaluation model.The standard abilities of each task characteristics are set according to the module classification.Combined with personal ability and task characteristics,the task-personnel matching degree model is established,and the matching degree is calculated.At last,we use the column difference method in the maximum difference method combined with constraints to find the optimal solution as the final task allocation plan.Taking S company as an example,this paper applies the new task allocation plan to the actual,analyzes and compares the implementation effect of the original task allocation plan based on business familiarity preference and the new task allocation plan from the perspective of time utilization,task personnel matching degree,personal ability improvement and team overall ability,and concludes that the task allocation plan considering the learning curve has higher time utilization and matching degree.It is more helpful to improve personal ability and the overall ability of the team,and improve the personnel’s coverage of the module.