Jackson networks and Markov processes for resource allocation modeling

摘要

Measuring the productivities of engineers is one difficult task in software engineering. The popular technique used to evaluate the quality and productivity of engineers during software development is that supervisors and managers have to monitor engineers and keep track their work activities every day. Managers might use their common sense that can be unfair and incorrect to evaluate performance of engineers. Hence two new models encouraged fault repair model and discouraged fault arrival model that we propose solve these issues. A closed form derived from encouraged fault repair model is used to measure the productivities of engineers, and a closed form derived discouraged fault arrival model to evaluate the performance of testers. The limitation the fault repair rate [Luong 2001] is a constant is removed in the new model, engineering resource configuration model that is a Jackson network model with Markov chain configuration. Factors causing workforce available or unavailable are included into the model.