A Study of Applying Agent Negotiation to Bank IT Project Managment

• Main Authors: Ching-Yi Wu, 吳靜怡
• Other Authors: 謝明哲
• Format: Others
• Language: zh-TW
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/5krbc8

碩士 === 國立臺東大學 === 資訊管理學系碩士班 === 97 === The information security and the system quality for the banking are extremely important. However, the financial IT industry in order to save the cost of training technical personnel expenditures, system outsourcing has been a trend. On present situation, bank IT has to establish a good consultative process and manage the project efficiently to have the opportunity to create the win-win situation. In the beginning of project, the plan proposed to be done must be drafted in consultation with the participants, so that the participants are willing to move forward toward common goals. When the project plan had been established, systems development also needs management. This is the project management duty for processing and coordination of emergency situation. Whether plan is beginning or proceeding, "project management" is an important issue. Therefore, it makes PM more important and necessary. However, project managers are not omnipotent in reality. On the premise of application and practicality, resources and suggestion provided by experts to solve the problem is project managers’ responsibilities, yet in progress of Project Management. In reality, it needs a lot of management attribute, such as manpower, cost, category, time and so on. These factors often change with environment. Our research established an Automated Negotiation Mechanism that progresses similar to human thinking process in project management on the basis of Multi-agent systems. In this study, we also build the system model to verify the feasibility of our consultation mechanism and genetic algorithm, and show the results of agent iv consulation simulation and the best special case combination suggested by the genetic algorithm when the negotiations fail. ※This study provides two core concepts: 1. When agent consultative mechanism endows agents with different negotiation behavior and tasks of PM, the automated negotiation system has become more user-friendly. 2. If the consultation fails, there will be a genetic algorithm trying to find the best combination. The purpose is to apply the agent in an integrated consultative mechanism, so that participants can consider their own interests in accordance with the properties to make adjustments flexible. For example, in the consultation, IT will adopt the most stringent requirements of system quality. In this regard, IT is impossible to make concessions. If all proposals cannot meet the participants’ needs, the genetic algorithm will be carried out to find the best combination according to the cost-effective value. Consequently, the best combination is used to offer participants further choice. In this research mentioned when consultation defeats, the system will try to find the best combination to give advice on the proposal, and to supply a record of the results to the agent system. By this way will accumulate and train thinking and decision-making capacity gradually, the model in the current research less mentioned. In the future may supply direction of the following research.