weakness – expert system

1. Require a lot of detailed knowledge
2. Restrict knowledge domain
3. Not all domain knowledge fits rule format
4. Expert consensus must exist
5. Knowledge acquisition is time consuming
6. Truth maintenance is hard to maintain
7. Forgetting bad facts is hard

Do this until problem is solved or no antecedents match

1. Collect the rules whose antecedents are found in WM.
2. If more than one rule matches use conflict resolution strategy to eliminate all but one
3. Do actions indicated in by rule "fired"

1. Level of Specificity

Maximum Specificity ( number of antecedents) match antecedents and choose the one with the most matches.

2. Physically order the rules. By doing this we can have subsets and we may also have preconditions
3. Order added to state description (WM)

{ recency ordering based on date }

4. Recency Ordering for rules
5. Data Ordering
6. Context Limiting
7. Execution Time
8. Fire All Application Rules

Simulation handout

R1 / XCON was created to solve problems

Spent 2 years to create the expert system

Stages

1. Check Order – missing/ mismatched pieces
2. Layout Processor Cabinets
3. Put boxes in input/output cabinets and components in boxes
4. Put panels in input/output cabinets
5. Layout floor plan
6. Indicates Cabling

OPS 5

• This is the first shell language
• R1/XCON 2500 rules

If context is layout and assign power supply then add appropriate power supply to order

1. Synthesis Systems

R1/XCON

Tends to be forward chaining

Often data driven

Often make use of breadth first search

This looks at all facts before proceeding

2. Analysis & Diagnostic

Given goal g

• find the set of rules that determine g
• if a set of rules does not equal empty set then

• choose rule R
• make R’s antecedent

• if new goal is unknown then backchain (ng)
• else

• until g is solved or S is equal to empty set
• else

end