Verification of Numeric Planning Problems Through Domain Dynamic Consistency

In Defence of Good Old-Fashioned Artificial Intelligence Approaches in Intelligent Transportation Systems

“…The knowledge of the considered problem to be solved is encoded in an abstract formal definition model, for the analysis, reasoning and manipulation by a planning engine in order to solve planning problems [43]. For any well-formed knowledge model there must be a sound interpretation of the dynamics [44], [45]. In other words, the knowledge model must provide a sound, accurate, and valid representation of the reality that is encoded.…”

“…Following the existing literature on explainable automated planning [15], we now focus on how the components above can support a question-based framework, that is the reference framework for explainable planning. In particular, we consider questions introduced in [15] that are relevant for the ITS domain, and that can provide valuable information to traffic experts and traffic engineers interacting with automated planning-based control systems: To better exemplify how the questions can be exploited by interacting with a planning-based ITS application, let us consider as a case study the example proposed in [45], based on traffic light optimisation in PDDL+ [28]. The knowledge model encodes the constraints and logic implemented by SCOOT systems to control a set of junctions, and a specific problem instance is provided by describing the initial network conditions (estimated or sensed) and a desired level of service.…”