A timeline, event, and constraint-based modeling framework for planning and scheduling problems

This paper presents a framework dedicated to the modeling of deterministic planning and scheduling problems. This framework is based on the notions of timelines (evolutions of state variables and of resource levels), events which impact timelines, and constraints on and between events. Building it has been motivated by the observation that many real planning and scheduling problems involve a finite set of events and that the main question when solving them is to assign values to event parameters (presence in the plan, position in the event sequence, precise date, specific type-dependent parameters) by satisfying a finite set of constraints on these parameters and optimizing some function of some of these parameters. This framework, we refer to as TECK for Timelines, Events, and Constraints Knowledge, makes it possible to express four kinds of knowledge that must be combined when planning and scheduling: knowledge about events (event presence and parameters), time (event positions and dates), state (state variable evolutions), and resources (discrete or continuous resource evolutions).