The utility of information and risk-taking fuzzy expert systems

✍ Scribed by J. J. Buckley; D. Tucker

• Publisher: John Wiley and Sons
• Year: 1988
• Tongue: English
• Weight: 941 KB
• Volume: 3
• Category: Article
• ISSN: 0884-8173
• DOI: 10.1002/int.4550030206

✦ Synopsis

In this article we employ utility theory to determine the new state of working memory, after a group of rules have been fired in parallel in a fuzzy expert system. This is argued to be analogous to using utility theory in economics to determine what best action to take in decision making under risk. A class of utility functions is described to compute the utility of information in working memory similar to computing the utility of wealth in economics. We discuss a memory update algorithm (fuzzy truth maintenance system) that will produce the unique undominated state of working memory after a group of rules have executed under the parallel modr ,f operation. The fuzzy expert system is called risk-averse when it uses this memory update algorithm. We call the system riskseeking, or risk-taking, when certain actions are allowed to operate outside the memory update algorithm. Experimenting with a risk-taking expert system is an exciting new idea which will exist in our new fuzzy expert system shell FESS 11.