Operations of AI production system
The operation of AI production system can thus be characterized as a search process in which rules are tried until some sequence of them is found. To solve a problem using a production system we must specify the database Rules
Control Strategy
Converting a problem statement into the above three components of a production system is the problem representation in AI. Once this is successfully done we only have problem solving which is just about searching in that problem space for the given goal state given the initial state.
A production system consists of productions (rules), a working memory of facts (database) and Control strategy is about an algorithm for producing new facts from old. As we have seen in the 8 puzzle example a rule becomes eligible to fire when its conditions match some set of elements currently in the working memory. As it matches left side the rule is fired and we have new facts, which is the facts that are generated on the right side of the rule.
A control strategy determines which of the several eligible rules fires next. It is important to realize the difference within production system and conventional computation.
Production System Versus Conventional Computation
Though both the systems use hierarchically organized programs, they have many differences .Irrespective of any local database, global database is accessible for everyone..
Here the communication takes place only between the databases but not between the rules.
Production system structure is modular; changes to any of the components made independently.
Using Conventional Computation in AI application is difficult, For any change in knowledge base would require extensive changes to the program