یک سیستم استدلال مبتنی بر مورد برای حل تعارض: طراحی و پیاده سازی
|کد مقاله||سال انتشار||تعداد صفحات مقاله انگلیسی||ترجمه فارسی|
|18753||2002||15 صفحه PDF||سفارش دهید|
نسخه انگلیسی مقاله همین الان قابل دانلود است.
هزینه ترجمه مقاله بر اساس تعداد کلمات مقاله انگلیسی محاسبه می شود.
این مقاله تقریباً شامل 10010 کلمه می باشد.
هزینه ترجمه مقاله توسط مترجمان با تجربه، طبق جدول زیر محاسبه می شود:
- تولید محتوا با مقالات ISI برای سایت یا وبلاگ شما
- تولید محتوا با مقالات ISI برای کتاب شما
- تولید محتوا با مقالات ISI برای نشریه یا رسانه شما
پیشنهاد می کنیم کیفیت محتوای سایت خود را با استفاده از منابع علمی، افزایش دهید.
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Engineering Applications of Artificial Intelligence, Volume 15, Issues 3–4, June–August 2002, Pages 369–383
A case-based reasoning system for conflict resolution called GMCRCBR is presented. The system is developed to provide real-time feedback to assist in the structuring and modeling of a conflict situation. A system architecture integrating the decision support system GMCR II, which implements a methodology called the graph model for conflict resolution, and the GMCRCBR system is developed. The issues of case representation, case storage, case retrieval, and case reuse are considered. Information on 104 conflict cases that have been analyzed using the graph model for conflict resolution is collected and stored. The system is first tested and verified using the case base and then the system's ability to retrieve similar cases is evaluated.
The study of conflicts and their resolution is one that transcends all areas of human endeavor. In fact, conflict seems to occur any time two or more human beings interact. The graph model for conflict resolution methodology (Fang et al., 1993) is developed for systematically studying real-world strategic conflicts. The basic components of a graph model for conflict are the decision-makers (DMs), the states, the possible state transitions and the decision makers who can unilaterally affect them, and the decision makers’ relative preferences over states. The graph model is currently implemented in practice by utilizing the decision support system GMCR II (Hipel et al (1997) and Hipel et al (2001); Peng, 1999). In GMCR II, a state is represented by a combination of options that are controlled by decision-makers. A large number of real-world conflicts have been documented and analyzed by utilizing the graph model for conflict resolution. During the initial modeling phase of a graph model, decision-makers, their options, the possible state transitions, and the decision-makers’ relative preferences over states must be identified. The identification of the decision-makers and their available options can be a time-consuming and difficult activity. An even more significant challenge is to ascertain the relative preferences that each decision-maker has for the feasible states. To overcome these challenges, it is useful to identify similar cases that have been analyzed using the graph model methodology. Case-based reasoning (CBR) utilizes the specific case information available as historical precedence for proposing solutions to current problems. The most important aspects of the existing cases are first stored and indexed. New problem situations are then presented and similar, existing cases are identified from the knowledge base. Finally, the previous problem solutions are adapted and the revised solutions are proposed for the current situation. The major objective of this paper is to present an interpretive CBR system for structuring and modeling a conflict. The system is proposed as a means to help overcome challenges in modeling a conflict with the graph model methodology by utilizing information provided in previous cases. The CBR system (GMCRCBR) is designed to transparently integrate with the existing decision support system (GMCR II), in order to provide real-time feedback during the modeling process. When providing advice to a client on how to strategically handle a conflict situation, an analyst or consultant would first have to develop in real-time a model of the dispute. More specifically, the analyst, based on information obtained from the client, would use the CBR system to enter a limited amount of information on a current case, including the involved decision-makers, as well as to retrieve similar cases that have been modeled and analyzed and to provide suggestions on the options that are available to them. Information from similar cases identified could assist the analyst and client in structuring and modeling the current conflict. As the model is developed and both the analyst and client will begin to understand the conflict better, intelligent feedback furnished by the system would be used to further refine the model that is later used as a basis for analyzing and generating strategic advice. It is anticipated that this CBR “guidance” will speed up and simplify the analysis process. Following an overview of CBR, a brief introduction to the graph model for conflict resolution is presented in Section 3 and two representative conflict cases are outlined in Section 4. The development of the CBR system for conflict resolution is described in Section 5 while the results of retrieval experiments for 104 cases are given in the next section.
نتیجه گیری انگلیسی
A CBR system has been developed for conflict situations that will integrate with existing analytical tools. This CBR system addresses a need to store conflict cases in a standardized format and to be able to retrieve this data in an efficient manner. The CBR system can also be utilized to help speed up the modeling process by filling in missing information for a conflict situation by retrieving information from similar archived cases. This feature will be especially useful for generating the available options for decision-makers, as this information is often difficult to obtain. The key contributions of the research presented in this paper are the design, implementation, and testing of a CBR system. During the design of the system, many different approaches to retrieval were investigated. The system demonstrated that it could retrieve similar cases for the user with error rates as low as 21.65% with a standard deviation of 24.43% for the nearest neighbor algorithm. The system was also able to retrieve useful options for the decision-makers 25.27% of the time with a standard deviation of 38.73%. These results could be improved as more cases representing the different problem domains are added to the case base. The next stage in development for the system will be to integrate the GMCRCBR system with GMCR II. The addition of the CBR functionality will provide an intelligent feedback mechanism to aid the user in modeling real-world conflicts using GMCR II. The integrated system will allow a user to more rapidly model a case using GMCR II through comparisons with similar cases that have been archived, and the suggestion of options that are applicable to the decision-makers. CBR is an exciting area where significant research is being conducted. The GMCRCBR system would benefit from research on any area of the CBR cycle. The case representation could be expanded to use abstraction hierarchies for the case features. Case templates could also be defined for each type of conflict or negotiation situation, rather than the more generic, single template used during this investigation. The storage methods could be expanded to support hierarchical data formats, and experiments can be conducted on better ways to arrange the data to provide faster retrieval times. The many retrieval methods offered in the GMCRCBR system could also be the topic of extensive research. The nearest neighbor algorithm could be expanded to assign different feature weights and selection based on the conflict type or another classification. Investigations could also focus on more appropriate methods for clustering the cases, which could provide significant improvements to the performance of the C4.5 algorithm. The performance of the system should also be re-evaluated as more cases are added to the case-base. Performance is expected to improve as more cases are entered into the system.