Publication: Value management-based alternatives ranking approach for automated negotiation
Date
2019
Authors
Mahmoud M.A.
Ahmad M.S.
Idrus A.
Journal Title
Journal ISSN
Volume Title
Publisher
Elsevier B.V.
Abstract
Ranking alternatives is considered one of the main challenges in automated negotiation by multi-agent systems. Several approached have been identified by literature to mitigate the challenges of alternatives ranking, however, these approaches overlooked the importance of value management principles in modeling. Value management emphasizes on two main criteria to make a value decision which are Cost verses Functions. In this paper, Value Management-based Alternatives Ranking Approach that includes cost and function criteria in the modelling. Two main approaches are proposed by literature to optimize the ranking which are interest-based ranking and weight-based ranking. However, the two approaches produce totally two different results for a particular problem. Alternative interest-based ranking is recommended by value-management; but it creates a huge conflict among members. While, Alternative ranking based-weight is straight-forward with low or no conflict, but it is influenced by power, such influencing reduces reliability of a decision. To tackle this problem, we combine the processes of the two approaches alternative interest-based ranking and weight to make value decision and at the same time reduce the conflict among members. A scenario from construction domain is selected to test the proposed approach in ranking the alternatives based on cost and function criteria. The result shows a promising approach that could assist in making value decision, in addition, the result also shows that interest-weight-based ranking reduce the conflict between parties in comparison with interest-based ranking and weight-based ranking. � 2019 The Authors.
Description
Automation; Decision making; Information systems; Information use; Alternative ranking; Automated negotiations; Ranking approach; Value management; Multi agent systems