Simulation-based effort estimation for COTS based software development

Abstract

Effort estimation has been considered as an important activity since the beginning of software engineering. It is important to direct the project planning and management of a software process by supporting decisions during the software process. It is equally important in the emerging software development paradigm CBSD (COTS Based Software Development). COCOTS(Constructive COTS model) is a developed algorithmic effort estimation model in CBSD. This research intends to improve the performance of the original COCOTS. The COCOTS model is a static, continuous and deterministic model. It gives the project manager or other decision maker a direction of effort factors and rough integration effort with a point estimation results. Considering the weaknesses in current effort estimation models and techniques (like COCOTS), the proposed approach is that of simulation-based integration effort estimation (referred to throughout this research as "COTSIM"). This simulation model is designed as a basis for the decision support in COTS software process. It is hypothesized that the new approach is more flexible than COCOTS by modeling the development process, uncertain factors, dynamic change of experience, individual COTS products, developers and their attributes. The experiment results demonstrate how the COTS selection, uncertainty, process parameters, experience of developers, and architecture technique influence the integration effort. The experiment also investigates how the resource allocation has impact on the duration of a project.