PDDL+ Publications Summary
A summary of the theoretical publications on PDDL+:
On Translation-Based Approaches from Discrete PDDL+ to Numeric Planning (ICAPS 2021): In this first article we sketched a semantic for PDDL+ under discrete semantics (commonly used in many PDDL+ planners but never formalised systematically). We have also proposed two translations, namely EXP and POLY, to translate a PDDL+ problem into a numeric problem (PDDL2.1, level 2). These translations are sound & complete w.r.t. the original PDDL+ problem under discrete semantics and the obtained PDDL2.1 problem can be addressed by using numeric planning systems.
A Sound (but Incomplete) Polynomial Translation from Discretised PDDL+ to Numeric Planning (AixiA 2021): In this article, we provide an alternative translation, namely POLY-, which is sound but incomplete in the general case. This translation is extremely compact and efficient for a subclass of PDDL+ problems.
The Power of Reformulation: From Validation to Planning in PDDL+ (ICAPS 2022): In this article, we have proposed a reformulation to re-present the validation problem of a PDDL+ plan into a PDDL+ (or numerical) planning problem. These reformulations can be exploited for both continuous and discrete validation depending on the reasoner employed.
A Practical Approach to Discretised PDDL+ Problems by Translation to Numeric Planning (JAIR, 2023): This journal article is a systematisation of the work presented at ICAPS 2021. The semantics for discrete PDDL+ are further refined.
Fixing Plans for PDDL+ Problems: Theoretical and Practical Implications (ICAPS 2023): In this article, we introduce a framework for fixing PDDL+ plans, namely Plan Fix Problem. In this framework, the goal is to repair an invalid plan according to a set of constraints that define the possible manipulations of the invalid plan. The theoretical properties are studied according to the type of constraints used and the Plan Fixing problem is tackled through reformulation.
A summary of the applicative publications based on PDDL+:
An Efficient Heuristic for AI-based Urban Traffic Control (MT-ITS 2023):
Automated Planning for Generating and Simulating Traffic Signal Strategies (IJCAI 2023):
PDDL+ Models for Deployable yet Effective Traffic Signal Optimisation (ICAPS 2024):