Backward chaining

Through a powerful technique known as backward chaining, PRPC allows a computation to advance even when the value of an input or parameter property is not available.

Declare Expression rules (Rule-Declare-Expressions rule type) and the Property-Seek-Value method work together to provide this capability. The system uses the internal dependency network of property relationships to develop inferences about how to obtain the missing property value.

Such computations are sometimes called goal-seeking, as the system reaches a final goal (finding a value for property X) by completing other goals (finding a value for properties that allow X to be computed).

Two standard flow actions VerifyProperty and VerifyPropertyWithListing provide examples of backward chaining.

	dependency network, forward chaining
	Declarative processing — Concepts and terms More about Declare Expression rules Property-Seek-Value method

Definitions