Prolog Incarnation

jakta-prolog-incarnation keeps the AgentSpeak(L)-inspired notion of beliefs and goals as Jason implements it:

• beliefs are Prolog facts and rules;
• goals are Prolog terms;
• triggers and guards work by unification, so a plan can match on(X, Y) and use X and Y in its body.

The logic engine comes from 2P-Kt, a Kotlin Multiplatform Prolog.

agent<PrologBelief, PrologGoal, Any> {
    embodiedAs { Any() }
    believes {
        +initialBelief { "parent"("abraham", "isaac") }
        +initialBelief { "parent"("isaac", "jacob") }
        +inferenceRule { "ancestor"(X, Y) impliedBy "parent"(X, Y) }
        +inferenceRule { "ancestor"(X, Y) impliedBy ("parent"(X, Z) and "ancestor"(Z, Y)) }
    }
    hasInitialGoals {
        !initialGoal { "introduce"("jacob") }
    }
    hasPlanLibrary {
        prologPlan {
            adding.goal {
                matchingGoal { "introduce"(P) }       // binds P = jacob
            } onlyWhen {
                satisfies { "ancestor"(A, P) }        // proves it on the beliefs, binds A
            } triggers {
                agent.print(P, " descends from ", A)  // jacob descends from isaac
            }
        }
    }
}

Why Prolog?

• Declarative knowledge. Inference rules derive new facts, like ancestor/2 above, so you don't have to store every consequence as a belief.
• Pattern-matching plans. A plan's trigger is a pattern, and one plan handles a whole family of goals.
• Jason compatibility. Belief sources ([source(bob)]) and KQML messaging work as they do in Jason, which makes porting AgentSpeak programs straightforward (see the Blocks World showcase).

In this section

• 2P-Kt integration and syntax: what JaKtA takes from 2P-Kt, its Kotlin DSL, which term builder to use where, annotations and sources, and reading values in Kotlin.
• Prolog plans: what prologPlan { } means and how variables flow through a plan.
• KQML messaging: what Jason-style tell, achieve, askOne, ... mean between agents.
• Caveats: the surprises to know about before you start.

To get things done, see the Prolog how-to guides: write guards and inference rules, define your Prolog vocabulary in Kotlin, ask and answer with KQML, turn perceptions into Prolog facts and port a Jason agent. For the full list of functions, see the Prolog Incarnation Reference.