24th International Conference on Flexible Automation and Intelligent Manufacturing
Permanent URI for this collectionhttps://hdl.handle.net/20.500.12588/880
Program and abstracts of the 142 papers accepted for presentation at the FAIM 2014 International Conference, held in San Antonio, Texas, and organized by the Center for Advanced Manufacturing and Lean Systems, University of Texas at San Antonio; Conference Chair: F. Frank Chen
Browse
Browsing 24th International Conference on Flexible Automation and Intelligent Manufacturing by Subject "Bayesian statistical decision theory"
Now showing 1 - 2 of 2
- Results Per Page
- Sort Options
Item A Bayesian network based decision support system(DEStech Publications, Inc., 2014) Chen, Wen-Hsin; Wang, Wei-Hua AndrewIn this highly dynamic environment, data and information are rapidly changing in time. For a decision support system, to capture continuum data and then to extract knowledge from data is in high demand. Our research work focus on developing a systematic approach for discovering causal relations among data and dynamic updating upon to the varying goals. Bayesian network is a powerful tool for representing knowledge and tackling the above inquires. In this research, the entropy-based searching mechanism designed on Bayesian network is adopted to learn the structure of a goal-specific mission. Link strength and connection strength are the measures for selecting links to be added and subtracted. However, these two measures are the relative measures instead of absolute thresholds. Initially, we designed a link/connection strength selecting method for possible adjacency matrices in tabu searching. In the second phase, we used a homogeneity test to ensure the quality of the learned structure. The results of this study will be an important stepping stone to attain the learning requirements of designing, building, and operating effective decision support systems under dynamic environment.Item Methodology for project risk assessment using Bayesian belief networks in engineering construction projects(DEStech Publications, Inc., 2014) Odimabo, O. O.; Oduoza, Chike F.Engineering construction projects commonly suffer from cost and time overruns, for most of the time because of uncertainties that are not carefully considered during bidding for contracts and budget project planning. These uncertainties place the project at risk of poor quality delivery and also not adhering to the time and budget schedule within the original contractual agreement. A clear focus on risk analysis and its management from the onset is essential to guide project planning and also to achieve optimal performance in construction projects. The research carried out here presents a risk assessment methodology based on the Bayesian belief network, which is an effective tool for knowledge representation and reasoning under conditions of uncertainty, structural learning procedure, combination of different source of knowledge, explicit treatment of uncertainty and support for decision analysis and fast responses for risk assessment. Bayesian belief network therefore, is a scenario planning tool suitable for project risk management because of its systematic and integrated process approach to the analysis of key risk factors affecting project delivery, with a view to predict the worst and best case scenarios and thereby guide project planning. The proposed methodology developed in this study is partly based on knowledge and experiences acquired from experts who are in a position to provide information on the sources of uncertainty, and the causes of uncertain condition with a view to generate optimal response strategies to support a successful project outcome.