在結構較不明確的決策問題上，決策者常諮詢各種資訊來源來改善決策品質，這種現象常發生於群體決策上，每個決策者以溝通來交換決策意見，並進而產生群體決策的共識解，因此，本研究將利用貝氏推論網路來架構多重智慧代理人系統進行決策時的協調溝通模式。將研究分兩階段，第一階段為群體成員各作獨立決策，以溝通來相互影響決策；第二階段為決策具有相依性的群體成員，協同作共同決策。在第一階段中，本研究提出如何在多個貝氏網路拓樸中找到一個最具溝通效率的樹狀結構，以加速成員間的意見交換；而第二階段則提出以貝氏網路來架構內部推論機制與外部決策相依現象的方法，並以資訊理論為基礎，最大化融合後概似機率的方式來作融合群組意見，以嘗試找出群體共識。最後，本研究於兩階段中，分別以供應商選擇與群體討論企業架站合宜性的範例，來說明發展的模式特性。由範例中可以看出，使用本研究的結構，可以讓代理人程式具有機率性推論能力，並藉由溝通與調適學習亦具有協同決策之功能，但本研究亦發現由於機率拓樸的相異、融合參數與調適參數的主觀設定之不同，可能造成意見的矛盾，而造成無法達成群體共識，與人類專家之溝通過程極其類似。

While making decisions in unstructured problems, decision makers usually consult various information sources to improve the decision quality. This phenomenon often exists while proceeding group decision making. Each decision member employs communication with others to exchange opinions, and then generate the consensus of group. Accordingly, this study tries to employ Bayesian Belief Networks (BBN) to construct the communication mechanisms while applying Multi-agents system to make decisions. It is classified into two stages: the first one is that group members make decisions independently, but communicate to generate some impact on others; the other is to construct a mechanism to facilitate dependent group members to make- decisions coordinately. At the first stage, it is proposed how to find the tree structure with the best communication efficiency among candidate BBN topologies; at the second one, it discusses how to employ BBNs to construct both the inward inference mechanism and the dependence of decision- making among groups. The proposed structure is based on the information theory and tries to generate group consensus by maximizing the fusing likelihood function. Finally, this thesis employs two experiments to explain the characteristics of these two stages, which are selecting best suppliers and discussing advantage of enterprise website with expert agents, respectively. Through the experiments, it could be observed that the proposed system could offer the abilities of probabilistic inference and coordinately decision- making with communication and adaptive learning. But with the results of the experiments, the group consensus is not always guaranteed while using different combination of probabilistic topology, pooling and adaptation parameters, since the contradictory opinions may occur. The characteristics of the proposed system are clarified to be very similar to the human beings’ communication.

Becker, Ann and Dan Geiger,” A Sufficiently Fast Algorithm for Finding Close to Optimal Clique Trees”, Artificial Intelligence, Vol.125, pp.3-17, 2001.

Bellazzi, Riccardo. and Alberto Riva, “ Learning Bayesian Networks Probabilities from Longitudinal Data”, IEEE Transactions on Systems, Man, and Cybernetics- Part A: Systems and Humans, Vol. 28, Iss.5, pp. 629-636, 1998.

Bryson, Noel,” Supporting Consensus Formation in Group Support Systems Using the Qualitative Discriminant Process”, Annals of Operation Research, Vol.71, pp.75-91, 1997.

Dumitrescu, Adrian, Bernd Gätner, Samuele Pedroni, and Emo Welzl,” Enumerating Triangulation Paths”, Computational Geometry, Vol.20, pp.3-12, 2001.

Gemela, Jozef,” Financial analysis using Bayesian networks”, Applied Stochastic Models in Business and Industry, Vol. 17, pp. 57-67, 2001.

Hajji, Hassan and Behrouz Homayoun Far,” Distributed Software Agents for Network Fault Management”, IEICE Transaction on Information and System, Vol.E83-D, Iss.4, pp.735-746, 2000.

Hendler, J.,” Making Sense Out of Agents”, IEEE intelligent Systems, Vol.14, Iss.2, pp.32-37, 1999.

Jacobs, Robert A.,“ Methods For Combining Experts’ Probability Assessments”, Neural Computation, Vol. 7, pp.867-888, 1995.

Jensen, Finn V., Bayesian Networks and Decision Graphs, Springer, New York, 2001.

Jensen, F.V., S.L., Lauritzen and K.G. Olesen,” Bayesian updating in causal probabilistic networks by local computations”, Computational Statistics Quarterly, vol. 4, pp.269-282, 1990.

Jiroušek, Radim and Nicholas Kushmerick,” Constructing probabilistic models”, International Journal of Medical Informatics, Vol. 45, pp.9-18, 1997.

Joo, Kwang Hyoun, Tetsuo Kinoshita and Norio Shiratori,“ Design and Implementation of an Agent- Based Grocery Shopping System”, IEICE Transaction on Information and System, Vol. E83-D, Iss. 11, pp.1940-1951, 2000.

Kullback, E. T., Information Theory and Statistics. Wiley, New York, 1959.

Kontkanen, P., P. Myllymaki, T. Silander, H. Tirri and P. Grunwald,” On predictive distributions and Bayesian networks”, Statistics and Computing, Vol.10, pp.39-54, 2000.

Kreng, Victor. B. and Chang, Chia-hua, “Bayesian Network Based Multiagent System-Application in E-Marketplace”, Computers & Mathematics with Application( To appear, 2003).

Lam, Wai and Fahiem Bacchus,” Learning Bayesian Belief Networks: An Approach Based on The MDL Principle”, Computational Intelligence, Vol. 10, No.3,pp.269-293, 1994.

Larrañaga, Pedro, Mikel Poza, Yosu Yurramendi, Roberto H. Murga, and Cindy M. H. Kuijpers,” Structure Learning of Bayesian Networks by Genetic Algorithms: A Performance Analysis of Control Parameters”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 18, No.9, pp.912-926, 1996.

Larson, Kate and Tuomas Sandholm,” Bargaining with limited computation: Deliberation equilibrium”, Artificial Intelligence, Vol. 132, pp.183-217, 2001.

Lau, H.C.W., S. K. Tso and J.K.L. Ho,” Development of an intelligent task management system in a manufacturing information network”, Expert Systems with Applications, Vol.15, pp.165-179, 1998.

Levy, William B. and Hakan Delic,” Maximum Entropy Aggregation of Individual Opinions”, IEEE Transactions on Systems, Man, and Cybernetics- Part A: Systems and Humans, Vol. 24, No. 4, pp.606-613, 1994.

Liu, H., M.X. Tang and J. Frazer,” Supporting learning in a shared design environment”, Advances in Engineering Software, Vol.32, pp.285-293, 2001.

Madejski, J.,” Agents as building blocks of responsibility- based manufacturing systems”, Journal of Materials Processing Technology, Vol. 106, pp.219-222, 2000.

McCutcheon, D. M. and F. Ian Stuart, “Issues in the choice of supplier alliance partners”, Journal of Operations Management, Vol.18, pp.279-301, 2000.

Morris, P.A., “ Combining expert judgment: A Bayesian approach”, Management Science, Vol.23, pp.679-693, 1977.

Mostaghimi, Mehdi, “Combining ranked mean value forecasts”, European Journal of Operational Research, Vol.94, pp.505-516, 1996.

Mostaghimi, Mehdi, “Bayesian Estimation of a Decision Using Information Theory”, IEEE Transactions on Systems, Man, and Cybernetics- Part A: Systems and Humans, Vol.27, Iss. 4, pp.506-517, 1997.

Pearl, J.,” Fusion, propagation, and structuring in belief networks”, Artificial Intelligence, Vol. 29, pp.261-296, 1986.

Pearl, J.,” Evidential reasoning using stochastic simulation of causal models”, Artificial Intelligence, Vol. 32, pp.245-257, 1987.

Rogers, T.J., Robert Ross and V.S. Subrahmanian,” IMPACT: A System for Building Agent Applications”, Journal of Intelligent Information System, Vol.14, Iss.2-3, pp.95-113, 2000.

Samarasooriya, Vajira N. S. and Pramod K. Varshney,” Sequential approach to asynchronous decision fusion”, Optical Engineering, Vol.35, No.3, pp.625-633, 1996.

Sanov, I.N.,”On the probability of large deviations of random variables”, Matematicheskii Sbornik , Vol. 42, Iss. 1, pp.11-44, 1957.

Seidenfeld, Teddy and Mark J. Schervish,“ Two Perspectives on Consensus for Bayesian Inference and Decisions”, IEEE Transactions on Systems, Man, and Cybernetics, Vol.20, Iss.2, pp.318-325, 1990.

Smith, Jim Q. and Álvaro E. Faria,” Bayesian Poisson models for the graphical combination of dependent expert information”, Journal of Royal Statistical Society, Vol.62. Iss.3, pp.525-544, 2000.

Soofi, E.S., “Capturing the intangible concept of information”, J. Amer. Statist. Assoc., Vol.89, Iss.428, pp.1242-1254, 1994.

Studeny, Milan and Remco R. Bouckaert,” On Chain Graph Models For Description of Conditional Independence Structures”, The Annals of Statistics, Vol.26, Iss.4, pp.1434-1495, 1998.

Turban, Efraim and Jay E. Aronson, Decision Support Systems and Intelligent Systems, Prentice Hall, New Jersey, 2001.

Wang, Haiqin, Denver Dash and Marek J. Druzdzel, “ A Method for Evaluating Elicitation Schemes for Probabilistic Models”, IEEE Transactions on Systems, Man, and Cybernetics—Part B: Cybernetics, Vol.32, No.1,pp.38-43, 2002.

Wangermann, John P. and Robert F. Stengel,” Principled negotiation between intelligent agents: a model for air traffic management”, Artificial Intelligence in Engineering, Vol.12, pp.177-187, 1998.

Xiang, Y., D. Poole, and M.P. Beddoes, “Multiply sectioned Bayesian networks and junction forests for large knowledge based systems”, Computational Intelligence, Vol.9, pp.171-220, 1993.

Xiang, Y., B. Pant, A. Eisen, M.P. Beddoes, and D. Poole, Multiply sectioned Bayesian networks for neuromuscular diagnosis”, Artificial Intelligence in Medicine, Vol.5, pp.293-314, 1993.

Xiang, Y.” A probabilistic framework for cooperative Multiagent distributed interpretation and optimization of communication”, Artificial Intelligence, Vol.87, Iss.1-2, pp.295-342, 1996.

Xiang, Y.,” Cooperative Triangulation in MSBNs Without Revealing Subnet Structures”, Networks, Vol.37, Iss.1, pp.53-65, 2001.