Agent Based Software Testing for Multi Agent Systems

Software testing starts with verification and validation and fulfills the requirement of the customer. Testing can be done by automation tool like Win runner, QTP or manually. If we talk about manual testing it takes lot of time and manpower also so nowadays we are using automation software. When we talk about automation testing so the cost of such kind of testing is very high so each company cannot afford. In this paper we are presenting agent based testing which is helpful for both kind of testing. Multi-Agent Systems (MAS) are characterized by autonomous and collaborative behaviors [1, 2]. Developing such systems is a complex process. As a result, a methodology for developing MAS is highly necessary. In this paper, a methodology using roles and ontology for such a purpose is presented [2]. The functionality of roles is estimated in the various phases of the MAS development. It is based on an emphasis on the properties and behaviors associated with each agent in MAS