L2Plan2: Learning Generalised Policies via Evolutionary Computation

L2Plan2 is an evolution-inspired system for inducing generalised planning policies from training data. A policy is here defined as a list of rules that specify which actions to be performed under which conditions. A policy is domain-specific and is used in conjunction with an inference mechanism that stipulates which rule to apply and which variable bindings to implement in order to formulate plans for problems within that domain.