The Complexity of Planning Revisited - A Parameterized Analysis
This work provides a more detailed complexity landscape for planning, explaining why some cases are simpler in practice than predicted, which is incremental for AI planning researchers.
The paper reanalyzes the computational complexity of planning subclasses in STRIPS and SAS+ using parameterized complexity analysis, showing that certain practical restrictions are tractable and a simple heuristic enables a partial-order planner to exploit this.
The early classifications of the computational complexity of planning under various restrictions in STRIPS (Bylander) and SAS+ (Baeckstroem and Nebel) have influenced following research in planning in many ways. We go back and reanalyse their subclasses, but this time using the more modern tool of parameterized complexity analysis. This provides new results that together with the old results give a more detailed picture of the complexity landscape. We demonstrate separation results not possible with standard complexity theory, which contributes to explaining why certain cases of planning have seemed simpler in practice than theory has predicted. In particular, we show that certain restrictions of practical interest are tractable in the parameterized sense of the term, and that a simple heuristic is sufficient to make a well-known partial-order planner exploit this fact.