Papers by Keyword: Case-Based Reasoning (CBR)

Paper TitlePage

Authors: Zi Qi Tang, Hui Ming Wu, Yi Wen Bian
Abstract: Because of low knowledge structural level in real tunnel engineering management, existing risk control methods (e.g., rule reasoning, decision tree inference) cannot completely meet the actual engineering demands. Considering the complexity of risk analysis of the tunnel engineering, this paper presents an aggregated risk control approach for tunnel engineering based on case library, i.e., representing case attribute modules using Frame representation method, reducing attributes using rough set technology, and simplifying case library by removing its redundancy in attribute similarity threshold.
Authors: Hong Yan Yang
Abstract: This paper applies case-based reasoning approach in university-industry problem-solving to solve problems emerged in knowledge flow when companies cooperate with universities. Solve the problems or conflicts in this domain are very important for companies, because these will decide whether companies could get competitive advantages or not. Capturing and preserving past experiences and knowledge of individuals in company by CBR will form corporate memory which facilitates the cooperation. Case structure and case features are presented in this paper.
Authors: Xian Yun Wang, Jian Qin Liu, Wei Guo
Abstract: Abstract For complex and difficult geology, it is difficult to design right cutters for TBM in the conventional ways. So the successful experiences and data accumulated are very useful in TBM disc cutters design. CBR is a reasoning methodology that exploits similar experienced solutions, in the form of past cases, to solve new problems. This paper proposes an AHP-Based CBR model that overcomes the difficulty of measuring experience for determining the relative weight of attributes by the analytic hierarchy process. By comparing, the model using the analytic hierarchy process was more accurate, reliable, and explanatory for solving new problems using experience from previous cases.
Authors: Rabiya Maharjan, Hong Chun Yuan
Abstract: Accurate prediction of aquatic product prices can improve the quality of business strategy of aquatic product market. Case-based reasoning (CBR) systems have long been intensively used in several areas of artificial intelligence. But it is difficult to cluster similar cases from case bases as there are uncertainties in knowledge representation, attribute description and similarity measures in CBR. To increase the efficiency and reliability of CBR, fuzzy theories have been combined with CBR. In this paper, fuzzy case-based reasoning (FCBR) has been developed to forecast the price of aquatic products.
Authors: Yun Dan Ren, Xiang Wang, Xiao Guang Rui
Abstract: Parting design, a classic work of strong experience and weak theory, has a direct impact on mold structure design and tooling. The method on the basis of case-based reasoning called CBR is one of the effective methods to solve problems of strong experience and weak theory. This paper proposes a frame of mold parting design of boat hull by UG on case-based reasoning, aimed at parting problems of telecontrolled boat hull.
Authors: Yi Jing Liu, Li Ya Chai, Jing Min Liu, Bo Wen Li
Abstract: The essential difference of cognitive radio from traditional radio lies in its ability to sense, learn and adapt to the environment. Recently, the research for cognitive radio has focused on the configuration problems of multi-objective optimization. However, in actual communication systems, the observable environment parameters are limited. Besides, the relationship between the system’s inputs and outputs is often complicated. Thus, Cognitive radio (CR) needs to understand and adapt to the environment through learning. To solve the problem mentioned above, a self-learning method for Cognitive radio decision engine based on CBR and Simulated Annealing is proposed. The simulation results show that the proposed method has the advantages of self-learning, multi-objective adaptation and rapid convergence.
Authors: Jing Kai Zhang, Peng Hui Li, Xiao Xiong Liu, Wei Guo Zhang
Abstract: Targeted at the faults of autopilot in fly-by-wire flight control system, a method based on integration of case and fault tree is proposed to quickly diagnose faults and provide the maintenance of autopilot system. In the process of diagnosis, according to the characteristics of fault information about autopilot system, the Build-in-test data was used to quickly search similar fault in the case base and accurately point out the fault location. For complex fault, the fault tree method is used to traverse to complete the reasoning and diagnosis procedure, and simultaneously add the results to the case base. The simulation results show the method is achieved.
Authors: Dong Min Zhang, W.H. Liao, X.S. Cheng, Y. Guo, S.X. An
Abstract: An enhanced product representation model was introduced, and it was specified with the sets of property, the sets of the multiplicity of modules, and the sets of relationships among them. A multi-case base corresponding complex product decomposition was established. Cases were defined and edited on customization template, which improved efficiency of developing multi-case base. For an example, a design tool for diesel engine development based on multi-case base was developed, and diesel engine product development could be more rapidly.
Authors: Zhuo Hu, Yong Zhao, Yang Chen, Dong Xiang
Abstract: G. Cortes Robles’ TRIZ (Russian acronym for “Theory of Inventive Problem Solving”)–CBR (Case-based reasoning) model has the capacity to accelerate conceptual design and to find inventive solutions. This paper describes how the TRIZ–CBR model can be enhanced by introducing the weight sensitivity. Of course, weight used to measure the relative importance of multiple attributes is a very important component of the similarity function, which has a great influence on the CBR results. By means of weight sensitivity analysis, the similar cases which possess high reference value among retrieved cases can be selected. A laundry drum product example was given to verify the validity of the model.
Showing 1 to 10 of 91 Paper Titles