Asian Research Thesis Index

Abstract

Software development success is based on understanding stakeholder needs and translates those needs into requirements. Requirement engineering is the process of discovering the purpose for which a system is intended. The success of any software product depends on how strongly the requirements are captured. Many daily activities require negotiations, often bilaterally. In such a kind of negotiation, two agents negotiate with each other to reach a consensus. In negotiation, different stakeholders are involved, and they might have contradictory requirements and different priorities. One of the reasons of failures of software is the conflict between stakeholders and developers. 
Different requirement engineering approaches are in practice but the Agent Oriented Requirement Engineering (AORE) focuses on objects and goals. The AORE models the requirements in terms of tasks and goals. Requirement engineering is mainly concerned with the identification of goals. REF allows segregate goals into hard and soft goals. These goals are then transformed into services and constraints. Then finally the responsibilities for the resulting requirements are assigned to agents such as human beings or software. 
Electronic negotiation system exists but if there is a hybrid negotiation (between human and computer) then this negotiation process becomes more complicated. The main obstacle in such kind of negotiation is incomplete information as the behavior of human is diverse by nature. Humans do make mistakes as they are influenced by the social and cultural issues. 
This research is intended to apply Requirement Engineering Framework (REF) and User Story Cards (USC) methodology to the human-agent bilateral negotiation to capture a complete set of requirements. By combining the strengths of these two approaches we presented how requirements can be captured efficiently. REF is a very simple model which has an effective graphical notation to represents objects. On the other hand USC allows to record requirements with consultation of customer representative. These cards clearly state the user view of requirements and states requirements very clearly. With the application of AORE, this thesis proposed a negotiation protocol that facilitates human-agent negotiation. The proposed protocol is run on a case study to show its implementation.

Add/Update Thesis