A decision support system utilizing a semantic agent

The adaptabilty, rapidity, and focus on high quality solutions offered by agile methodology have lead to a paradigm shift in the software development process in many enterprises. Agile methodology is iterative in nature, with each iteration i.e. timebox lasting 2-6 weeks. Iterations involve small teams comprising 9-19 developers working through the entire software development life cycle. Agile methodology works on two basic principles. The first being regular adaptation to changing circumstances and the second -focus on technical excellence and good design and high quality code. The first principle accommodates that tasks in an agile project cannot be predicted more than a week in advance. Thus the need arises for project teams to incorporate experts in the problem domain, such that they are better equipped to handle changes rapidly. However this methodology has been criticised as it may not bring about the benefits intended by the second principle unless practised by skilled programmers, who can create high quality code. Hence a project manager should be equipped with a highly skilled team. We propose the utilization of a semantic agent, which will act on behalf of the project manager and suggest experts based on a set of parameters. Our semantic agent is based on a semantic matching algorithm. This algorithm utilizes an ontology based similarity framework to make recommendations and suggest training paths to satisfy the requirements of the project manager. The agent uses this algorithm to recommend employees based on their expertise, past experience and availability. Further, based on recommendations made by the agent we classify employees as experts and non experts and suggest knowledge transfer methods to upgrade their skills.