A prolog implementation of pattern search to optimize software quality assurance

Abstract

Quality Assurance (QA) is a critical factor in the development of successful software systems. Through the use of various QA tools, project managers can ensure that a desired level of performance and reliability is built into the system. However, these tools are not without cost. Project managers must weight all QA costs and benefits for each development environment before weigh all QA costs and benefits for each development environment before establishing an allocation strategy. The development of a system dynamics model has provided project managers with an automated tool that accurately replicates a project's dynamic behavior. This model can be used to determine the optimal quality assurance distribution pattern over a given project's life cycle. The objective of this thesis was to enhance a prototype expert system module that interacts with the system dynamics model for determining QA effort allocation schemes. The new module uses a pattern search algorithm to derive an optimal distribution scheme from a given set of project parameters. This system not only resolves all limitations discovered in the prototype model but also achieved significant reductions in total project cost.