AI Production Systems

An AI production system is a system that uses rules to make decisions.

The major elements in an AI production system are a global database, a set of production rules and a control system.

Global database: A central data structure used by an AI production system to store information about the world.
Production rule: A rule which operates on the global database. Each rule has a precondition that is either satisfied or not by the global database.
- If the precondition is satisfied, the rule can be applied, else not.
Control system: A system that chooses which applicable rule should be applied and ceases computation when the termination condition is satisfied on the global database.

Solving Problems with AI Production Systems

To solve a problem using an AI production system, we must specify the global database, a set of production rules and the control strategy. Transforming a problem statement into these three components in a production system is known as representation problem in AI.

A control strategy involves selecting rules and keeping track of those sequences of rules already tried, and the databases they have produced for a production system.

Basic Procedure of AI Production Systems

The basic procedure for production system algorithm for solving a problem (for example, the 8 puzzle problem) can be written in non-deterministic form as follows:

DATA \(=\) initial database
while DATA satisfies the termination condition, do:
begin
select some rule \(R\) from the set of rules that can be applied to DATA
DATA \(=\) result of applying \(R\) to DATA
end

Classification of Control Strategies

Control Strategies:

Irrevocable Control Strategies
- Hill Climbing Algorithm
Tentative Control Strategies
- Backtracking control strategy
- Graph search control strategy

Irrevocable Control

In this control strategy, an applicable rule is selected and applied irrevocably without provision for reconsideration later (no Backtracking is allowed).

An example of an algorithm that directly makes use of the irrevocable control strategy is the Hill Climbing Algorithm.

Tentative Control

In tentative control regime, an applicable rule is selected (either arbitrarily or perhaps with a good reason). The rule is applied, but provision is made to return later to this point in the computation to apply some other rule.