Abstract:
Electronic negotiations are becoming an important research subject in the area of electronic commerce. Decision analysis and especially multiattributive utility theory play an important role for the support of electronic negotiations. Decision theory is a well-tested approach to helping a user choose from a set of alternatives when the user's preference structure is known. The preferences are usually represented as a utility function on the set of alternatives such that the user prefers an alternative exactly when it has higher utility. In cases where the utility function is unknown or partially known, it must be elicited from the user in order for decision theory to be applicable, and this can often be a time consuming process that has unknown payoff for the user. Understanding and representing a user's preferences accurately is a crucial element of most decision support systems, since users have a bounded rationality, such complex processes will never be strictly linear nor rational. Therefore, there are a number of challenges that NSS has to meet. This paper discusses the challenges and evaluates different methods for preference elicitation with respect to these challenges. To choice the appropriate preferences elicitation method and to guess the opponent's preferences can overcome the specific challenges of preference elicitation in NSS which will be suitable for dynamic ecommerce.