2.2. Rete Algorithm

When using Rete00, the root node is where all objects enter the network. From there, it immediately goes to the ObjectTypeNode.

The ObjectTypeNode helps to reduce the workload of the rules engine. If there are several objects and the rules engine tried to evaluate every single node against every object, it would waste a lot of cycles. To make things efficient, the ObjectTypeNode is used so the engine only passes objects to the nodes that match the object's type. This way, if an application asserts a new Account, it won't propagate to the nodes for the Order object.

In JBoss Rules, an object which has been asserted will retrieve a list of valid ObjectTypesNodes via a lookup in a HashMap from the object's Class. If this list doesn't exist it scans all the ObjectTypeNodes finding valid matches which it caches in the list. This enables JBoss Rules to match against any Class type that matches with an instanceof check.

AlphaNodes are used to evaluate literal conditions. When a rule has multiple literal conditions for a single object type, they are linked together. This means that if an application asserts an Account object, it must first satisfy the first literal condition before it can proceed to the next AlphaNode.

AlphaNodes are propagated using ObjectTypeNodes.

JBoss Rules uses hashing to extend Rete by optimizing the propagation from ObjectTypeNode to AlphaNode. Each time an AlphaNode is added to an ObjectTypeNode it adds the literal value as a key to the HashMap with the AlphaNode as the value. When a new instance enters the ObjectType node, rather than propagating to each AlphaNode, it can instead retrieve the correct AlphaNode from the HashMap, thereby avoiding unnecessary literal checks.

BetaNodes are used to compare two objects and their fields. The objects may be the same or different types.

Alpha memory refers to the left input on a BetaNode. In JBoss Rules, this input remembers all incoming objects.

Beta memory is the term used to refer to the right input of a BetaNode. It remembers all incoming tuples.

When facts enter from one side, you can do a hash lookup returning potentially valid candidates (referred to as indexing). At any point a valid join is found, the Tuple will join with the Object (referred to as a partial match) and then propagate to the next node.

A LeftInputNodeAdapter takes an Object as an input and propagates a single Object Tuple.

Terminal nodes are used to indicate when a single rule has matched all its conditions (that is, the rule has a full match). A rule with an 'or' conditional disjunctive connective will result in a sub-rule generation for each possible logically branch. Because of this, one rule can have multiple terminal nodes.

Node sharing is used to prevent unnecessary redundancy. Because many rules repeat the same patterns, node sharing allows users to collapse those patterns so they don't have to be re-evaluated for every single instance.

The following two rules share the first pattern, but not the last:

rule
when
    vehicle( $car : name == "car" )
    $driver: Driver( typeCar == $sedan )
then
    System.out.println( $driver.getName() + " drives sedan" );
end

rule
when
    Vehicle( $sedan : name == "sedan" )
    $driver : Driver( typeCar != $sedan )
then
    System.out.println( $driver.getName() + " does not drive sedan" );
end