![]() |
|||
![]()
|
![]() |
![]() |
![]() |
2.10. GEOX: A KNOWLEDGE-BASED DECISION SUPPORT SYSTEM FOR MINERAL EXPLORATIONStabell et al. GeoX shares some of the same technical features as the SPECTRUM, META/LOG, and Contouring Assistant. It is a commercial system that has generated significant sales during the 1990s. GeoX is another case of a commercial success that is built on a hybrid approach, object-oriented technology, and multiple knowledge sources set in a blackboard type of fashion. In addition to this, it contains some characteristics that are particular to that system. GeoX was built to support the exploration process, primarily at the play level, but later versions address prospect-specific issues. Recent extensions also include economic aspects. The decision process thus supported is often referred to as "full cycle." The key features of the original GeoX (Stabell 90; Mikkelsen 92) can be listed as follows:
Acknowledging the fact that decision-making addresses a vast amount of issues and data GeoX aims at supporting the explorationist through advice and critique rather than pursuing the exploration objectives on its own. In fact, the GeoX project recognized the fact that real value lies in harmonizing the contributions of many decision-makers. With a hit rate of approximately 10% in areas such as the North Sea, human geologists are performing at a level that oil company management and government authorities deem grossly unsatisfactory. The philosophy behind GeoX thus became two-folded: to organize the decision-making process in a structured manner so that work going on in different areas could be compared; and to concentrate not so much on the result of the decision-making like ordinary expert systems in the domain, but toward its rationale. By mobilizing expertise to standardize and structure the argumentation process of individuals, the work performance could be enhanced. Hence, a structured and knowledge-based justification tool was added to GeoX. This tool determines suitable justification templates for the type of arguments that the user applies. Increased transparency of the decision-making, comparison of argumentation, and biases as well as impoved sharing and re-use of decision strategies could be achieved. GeoKnowledge later published the first initiative results from an empirical study (Mikkelsen 96) where the GeoX's approach was extensively tested. Six explorationists were given a modified Southeast Asian play. They had 2 hours of clock time for understanding the goal of the system and the purpose of the evaluation, and for producing an interpretation by interacting with GeoX. The increased conformation of the decision process introduced by the structured justifications imposed by GeoX did not reduce the effectiveness of the explorationists. The benefits of making it possible to compare each person's approach to the qualitative aspects of the play analysis are obvious. It enables the management of an efficient group process involving multiple experts and a vast amount of experience.
The structured justification approach resides on a model-based description of the play implemented by means of objects (see Figure 4). This defines a context for the user in which he enters facts and hypotheses about the play. Given the basic input quantities and contextual descriptions entered by the user, the system uses an estimation tool that produces a report that links calculated output like recoverable reserves to risk probabilities and the qualitative rationale produced together with the explorationist. GeoX was implemented by means of an in-house tool called K2. K2 is a knowledge representation environment for Smalltalk-80. The most recent version runs under VisualWorks 2.5.1.
|
![]() |
|
Use of this site is subject certain Terms & Conditions. Copyright (c) 1996-1999 EarthWeb, Inc.. All rights reserved. Reproduction in whole or in part in any form or medium without express written permission of EarthWeb is prohibited. Please read our privacy policy for details. |