Brought to you by EarthWeb
IT Library Logo

Click Here!
Click Here!

Search the site:
 
EXPERT SEARCH -----
Programming Languages
Databases
Security
Web Services
Network Services
Middleware
Components
Operating Systems
User Interfaces
Groupware & Collaboration
Content Management
Productivity Applications
Hardware
Fun & Games

EarthWeb Direct EarthWeb Direct Fatbrain Auctions Support Source Answers

EarthWeb sites
Crossnodes
Datamation
Developer.com
DICE
EarthWeb.com
EarthWeb Direct
ERP Hub
Gamelan
GoCertify.com
HTMLGoodies
Intranet Journal
IT Knowledge
IT Library
JavaGoodies
JARS
JavaScripts.com
open source IT
RoadCoders
Y2K Info

Previous Table of Contents Next


2.4. LITHO

Bonnet et al.
Schlumberger/LRI/University Paris-Sud, 1982

The purpose of LITHO was to interpret oil-well measurements (Bonnet 83). Inherent in this is the need to elucidate curves measuring physical properties of rocks. Apart from this, the interesting thing about LITHO was the fact that it was built by means of the first expert system shell created, EMYCIN. As such, it was based on the original MYCIN concept using rules. LITHO contained several hundred production rules that contained knowledge of both petrography and paleontology.

The EMYCIN shell created a structure for the final system together with a fixed type of rule schemata like that used in MYCIN. A well-defined interface enabled the knowledge engineer to concentrate on the knowledge issues rather than on software engineering. For a suitable type of problem, this could be very efficient. However, a shell like this often proved too narrow for many problems and the knowledge engineer could often find himself fighting the limits of the shell rather than concentrating on the cognitive issues. But LITHO proved that it was possible to build a substantial system using a standard off-the-shelf shell, although problems like the one indicated were encountered during the building process.

2.5. META/LOG: A KNOWLEDGE-BASED EXPERT SYSTEM FOR LOG ANALYSIS

SPECTRUM 2000 Mindware
Denver, Colorado, 1987

Throughout the era of expert systems, two application ideas have prevailed. One is the benefits of efficient knowledge sharing; the other is that of separating a body of knowledge from a single person and make it available independently of the human. The latter can be perceived as a manager's dream. Seen from the point of view of an oil company, it can be difficult to accept the fact that their primary source of income and asset making and the employment of as many as 60,000 to 70,000 employees are dependent on a few hundred explorationists. A top score explorationist may take 15 to 20 years to develop his skills and expertise. Still his services and advice are unreliable and expensive.

DIPMETER ADVISOR had succeeded in isolating a type of exploration knowledge and made it available to a broader audience. SPECTRUM 2000 Mindware sought to exploit commercial benefits through its endeavors with the META/LOG system, which they claimed incorporated more than 25 years of log analysis experience. SPECTRUM 2000 Mindware was basically a consulting firm that was one of the first within the geological domain and possibly within the whole world of expert system development to seek commercial opportunities through sales of computer-based knowledge. Instead of relying on sales of knowledge at an hourly rate and databased computer systems, SPECTRUM 2000 Mindware pursued repetitive sales of their log analysis knowledge through the building and marketing of the META/LOG system.

In this discourse, we will use SPECTRUM 2000 Mindware mainly as a case of commercialization of expert systems, although the developers of META/LOG were bold enough to apply some very novel techniques that are interesting in their own right.

META/LOG was launched as a knowledge-based expert system for log analysis (SPECTRUM 87), although its scope was much broader than many of its successors. It contained economical implications of the log interpretations that had not been part of comparable knowledge bases at the time. It produced conclusions on such things as reservoirs, recoverable reserves, productivity, and potential cash flow.

When it was launched in 1987, META/LOG was part of a suite of programs including databases on well history, traditional log and core analysis processors, graphics, and a system for oil exploration economics. The system itself featured a set of modules:

  • A database containing raw log data, default analysis parameters and a mathematical model for analyzing the data
  • A knowledge acquisition module for determining the local geologic setting and preferences for the user
  • A rule base for determining the best log analysis method, based on the available data, borehole conditions, and the expected rock and fluid type
  • A knowledge base containing rock, fluid, and reservoir properties for the most common reservoir types
  • A rule base for determining the appropriate parameters for the selected method and the expected rock and fluid type

In order for all these components to work together, the developers applied a novel architecture at the time called the "blackboard." Each module would communicate through the blackboard, posting its hypotheses and facts there. The blackboard would thus hold the current status of the shared problem-solving effort. A meta-module would assess the posted goals and determine any changes in the problem-solving strategy. The blackboard paradigm originated with the Hearsay programs at Carnegie-Mellon University (Lesser 77) and was a way of making individual knowledge bases communicate and cooperate to reach a solution that they could not produce on their own.

The basic control paradigm applied beyond this was secured by the use of an inference engine applying the various contributions in a forward-type of manner. In addition, an audit trailer was included in order to monitor the problem-solving process. The developers realized the need for allowing the user stay in control. Users of the traditional expert system approach could not control the problem-solving process. The system itself made the initiatives. It prompted and expected an answer. Once finished, it would launch a nonnegotiable answer. META/LOG included a manual override mode that could fine-tune the results produced by the system or bypass them. An important aspect, considering that this was a purely commercial initiative, was the fact that META/LOG did not require sophisticated hardware nor any particular software considerations. It was meant for the IBM-PC or compatibles and came in two versions, implemented on top of Lotus 1-2-3 or Symphony. The entire suite of programs was marketed at a price of $950.


Previous Table of Contents Next

footer nav
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.