Chaining › Forward Chaining

Forward Chaining

Sometimes you want your application to accept all input data and then start firing applicable chains of rules. The process can continue until all the consequences of the input data have rippled through the knowledge base.

The result of executing your application is a set of attribute values on which business decisions can be based. The actions of some rules may also execute methods to perform such business actions as modifying a database, generating notification letters, or contacting a credit bureau. This type of rule processing is called forward chaining and has the following characteristics:

• It is data-driven reasoning
• The inference engine generally considers rules in the order posted
• The inference engine can determine the full ripple effect of input data

  (You can have the inference engine exhaust all rules before it terminates.)