Publication Date


Degree Program

Department of Mathematics and Computer Science

Degree Type

Master of Computer Science


Using an expert system to make decision making more reliable has been well studied and implemented over the years. For effective use, both data-driven questions (forward chaining) and goal-driven questions (backward chaining) need to be supported. Similarly, an avenue to update rules in the system as and when they change without major recompilation should be available. In this thesis we present an expert system framework that can help small water system operators make informed decisions regarding compliance with various EPA rules that may apply to them. To support both types of questions mentioned earlier, the system incorporates two expert system shells: JESS for answering data-driven questions such as "This is my reading for sample X. What needs to happen next?" and MANDARAX for goal-driven questions such as "We want to be compliant with the Total Coliform Rule. What do we need to do?" To make sure that rules are consistent and to support a straightforward rule-updating process, we use a native xml database to store the rules. All the rules are in XML format which ensures better symbiosis with other tools that support XML and allows one set of rules to be used for both JESS and MANDARAX.


Computer Sciences