UNIQUE TECHNOLOGIES

Products

CEM

Climate Change

Energy and
    Environmental Decision
    Support Systems

Unique Technologies

Papers:

Beyond Data Mining

Process Management

Distributed/Approximate
     Reasoning

Problems with Traditional Technologies

     Experience shows (Control Engineering, February 1995, p.12) that the key barriers to use of intelligent technologies in industry are:

• Users hestitate to accept recommendations from PI systems when there is no clear explanation of how the recommendation was arrived at.

• Changes to the process or production that require PI systems to be upgraded (at a substantial cost)!

    
     Older technologies like Expert Systems or Neural Networks only address one of the barriers listed:

• Expert Systems offer explanations of their recommendations but are expensive to upgrade

• Neural Networks are relatively easy and inexpensive to upgrade, but do not offer explanations (transparency) of how they arrive at the recommendations.

A Radically Different Approach

     REDUCT Decision and Process Suites offer a radically different approach to decision support and optimization by incorporating users into the decision making process.  They provide the user with all background information and results for alternative solutions so an informed choice can be made.  This approach emphasizes helping users in their actions by making all the decisions and recommendations transparent and easy to understand.  This greatly reduces the risk of a major mistake due to “misbehavior” of the process control models.

     The knowledge discovery tools in  the REDUCT Decision Suite provide the means to extract knowledge from and develop a decision support and optimization system more quickly and less expensive than in traditional expert systems.  The use of approximate reasoning methods reduces the needs for frequent knowledge base updates by allowing decisions to be made even though the present does not exactly replicate the past.  However, the decision support and optimization system will indicate when the current knowledge base is no longer adequate.   When this happens, the system is relatively easy and inexpensive to update.

Summary

     REDUCT Decision and Process Suites provide unique solutions, which are transparent and easy to update.  Specifically they address issues and requirements for:

1. knowledge discovery methods that can handle incomplete, inconsistent and/or imprecise data.

2. decision support systems that can process situations where data are missing and/or ambiguous.

3. adaptive optimization strategies that represent the current dynamics of the system.

4. optimization systems that clearly describe control states and the reasoning used to reach the conclusions and recommendations.

5. rapid and inexpensive updating of process control knowledge.

REDUCT & Lobbe Technologies Inc.
P.O. Box 800,  186 - 8120 No.2 Road., Richmond, BC,  Canada  V7C 5J8
ph: (604) 275-3711   fax: (604) 275-3711  email: dispatch@reduct.com