Genie Extension Study

With the social progress and technological development, people in their daily lives will be faced with the uncertainty of decision problems. The influence diagram is based on uncertain information and the graph model for solving complex decision problems. Influence diagram could be used its node and arc characteristics and information content characterization preferences, probability estimates, options and state, full use of the conditional independence relationships between variables, so that nodes do not have to rely on all the former trend, a clear distinction between timing information and the probability of the corresponding joint probability density expansion from the natural order of the decision makers thinking, easy to decision-makers understand the intuitive graphical comprehensive insight into the problem, and standardized decision-making problems more effectively.However, with the expanded scale of the uncertainty of decision-making problems, a simple artificial influence diagram model of decision-making and utility computing is too much trouble, large calculation, time-consuming, and error-prone, so the researchers developed a number of auxiliary decision-making tools such as GeNIe, Analytica, Decision Manager, and so on. GeNIe is supporting software developed by the University of Pittsburgh Decision Systems Laboratory. Using GeNIe could compute traditional influence diagram utility, and find the optimal strategy. However, due to the decision-making environment and the problem is complex and diverse, there is no any kind of modeling tools to be generally applicable to all decision-making problems. Researchers on the influence diagram, a variety of expansion, for example, with the weight of the qualitative influence diagrams and with interval parameters influence diagrams and so on. These new influence diagram model GeNIe software has been unable to apply. Based on this phenomenon extends the right GeNIe software, an increase in the GeNIe with weights qualitative influence diagrams and with interval parameters influence diagrams of two models, and were added to the evaluation algorithm. This paper first introduces the basic concepts of the Bayesian network, affecting the basic concepts of graph, with the weight of the qualitative influence diagram of the basic concepts and decision-making algorithm with the basic concepts of interval parameters and decision-making algorithm.Secondly, the Genie and SMILE basic functions and use, and study how to use the SMILE development GeNle.Thirdly, to use GeNIe background framework SMILE extend GeNIe in GeNIe two influence diagrams,the qualitative influence diagrams with weight and influence diagrams with interval parameters both influence diagrams model.Finally, the experiments show extended results, and present work and a summary and outlook for future work.