SAT-Based Parallel Planning Using a Split Representation of Actions (original) (raw)

A Compact and Efficient SAT Encoding for Planning

2008

In the planning-as-SAT paradigm there have been numerous recent developments towards improving the speed and scalability of planning at the cost of finding a step-optimal parallel plan. These developments have been towards: (1) Query strategies that efficiently yield approximately optimal plans, and (2) Having a SAT procedure compute plans from relaxed encodings of the corresponding decision problems in such a way that conflicts in a plan arising from the relaxation are resolved cheaply during a post-processing phase. In this paper we examine a third direction of tightening constraints in order to achieve a more compact, efficient, and scalable SAT-based encoding of the planning problem. For the first time, we use lifting (i.e., operator splitting) and factoring to encode the corresponding n-step decision problems with a parallel action semantics. To ensure compactness we exploit reachability and neededness analysis of the plangraph. Our encoding also captures state-dependent mutex constraints computed during that analysis. Because we adopt a lifted action representation, our encoding cannot generally support full action parallelism. Thus, our approach could be termed approximate, planning for a number of steps between that required in the optimal parallel case and the optimal linear case. We perform a detailed experimental analysis of our approach with 3 state-of-the-art SAT-based planners using benchmarks from recent international planning competitions. We find that our approach dominates optimal SAT-based planners, and is more efficient than the relaxed planners for domains where the plan existence problem is hard.

Parallel Encodings of Classical Planning as Satisfiability

Lecture Notes in Computer Science, 2004

We consider a number of semantics for plans with parallel operator application. The standard semantics used most often in earlier work requires that parallel operators are independent and can therefore be executed in any order. We consider a more relaxed definition of parallel plans, first proposed by Dimopoulos et al., as well as normal forms for parallel plans that require every operator to be executed as early as possible. We formalize the semantics of parallel plans emerging in this setting, and propose effective translations of these semantics into the propositional logic. And finally we show that one of the semantics yields an approach to classical planning that is sometimes much more efficient than the existing SAT-based planners.

Planning as satisfiability: parallel plans and algorithms for plan search

Artificial Intelligence, 2006

We address two aspects of constructing plans efficiently by means of satisfiability testing: efficient encoding of the problem of existence of plans of a given number t of time points in the propositional logic, and strategies for finding plans given the formulae representing these formulae for different values of t.

SAT-based planning in complex domains: Concurrency, constraints and nondeterminism

Artificial Intelligence, 2003

Planning as satisfiability is a very efficient technique for classical planning, i.e., for planning domains in which both the effects of actions and the initial state are completely specified. In this paper we present C-SAT, a SAT-based procedure capable of dealing with planning domains having incomplete information about the initial state, and whose underlying transition system is specified using the highly expressive action language C. Thus, C-SAT allows for planning in domains involving (i) actions which can be executed concurrently; (ii) (ramification and qualification) constraints affecting the effects of actions; and (iii) nondeterminism in the initial state and in the effects of actions. We first prove the correctness and the completeness of C-SAT, discuss some optimizations, and then we present C-PLAN, a system based on C-SAT. C-PLAN works on any C planning problem, but some optimizations have not been fully implemented yet. Nevertheless, the experimental analysis shows that SAT-based approaches to planning with incomplete information are viable, at least in the case of problems with a high degree of parallelism.

A novel constraint model for parallel planning

A parallel plan is a sequence of sets of actions such that any ordering of actions in the sets gives a traditional sequential plan. Parallel planning was popularized by the Graphplan algorithm and it is one of the key components of successful SAT-based planers. SAT-based planners have recently begun to exploit multi-valued state variables-an area which seems traditionally more suited for constraint-based planners-and they improved their performance further. In this paper we propose a novel view of constraint-based planning that uses parallel plans and multi-valued state variables. Rather than starting with the planning graph structure like other parallel planners, this novel approach is based on the idea of timelines and their synchronization.

Faster probabilistic planning through more efficient stochastic satisfiability problem encodings

2001

The propositional contingent planner ZANDER solves finitehorizon, partially observable, probabilistic planning problems at state-of-the-art-speeds by converting the planning problem to a stochastic satisfiability (SSAT) problem and solving that problem instead (Majercik 2000). ZANDER obtains these results using a relatively inefficient SSAT encoding of the problem (a linear action encoding with classical frame axioms). We describe and analyze three alternative SSAT encodings for probabilistic planning problems: a linear action encoding with simple explanatory frame axioms, a linear action encoding with complex explanatory frame axioms, and a parallel action encoding. Results on a suite of test problems indicate that linear action encodings with simple explanatory frame axioms and parallel action encodings show particular promise, improving ZANDER’s efficiency by as much as three orders of magnitude.

SAT-based Weighted Planning

2008

Automated planning is one of the most important problems in artificial intelligence. We present a new refinement of the classical planning algorithm that formulates the planning problem as a satisfiability problem. Compared with previous techniques, the solution of the planning problem is identified using the number of truth assignments of the corresponding propositional formula and their actions' utilities. Our approach eliminates backtracking and supports efficient planners that consider additional subformulas without the need to recompute solutions for previously provided subformulas. The experimental results show that our approach can help existing SAT-based state-of-the-art planners to find the solution plan more efficiently. * This paper is an improved and unified presentation of conference papers [2] and . It contains a new section (Section 5) and complete proofs for the theoretical results (e.g., Lemma 4.1, 4.2, and Theorem 4.1), new definitions and more comprehensive experimental results.

A hybrid bit-encoding for SAT planning based on clique-partitioning

AIP Conference Proceedings, 2017

Planning as satisfiability is one of the most efficient ways to solve classic automated planning problems. In SAT planning, the encoding used to convert the problem to a SAT formula is critical for the performance of the SAT solver. This paper presents a novel bit-encoding that reduces the number of bits required to represent actions in a SATbased automated planning problem. To obtain such encoding we first build a conflict graph, which represents incompatibilities of pairs of actions, and bitwise encode the subsets of actions determined by a clique partition. This reduces the number of Boolean variables and clauses of the SAT encoding, while preserving the possibility of parallel execution of compatible (non-neighbor) actions. The article also describes an appropriate algorithm for selecting the clique partition for this application and compares the new encodings obtained over some standard planning problems. CLASSICAL AUTOMATED PLANNING AS BOOLEAN SATISFIABILITY Automated Planning is a part of Artificial Intelligence that states the problem of selecting a course of actions to reach a goal. Classical automatic planning employs a planning model that on a domain description-typically an initial state and a goal state-and a set of actions that can change the current state. Classical domains have some restrictions: they are fully observable, deterministic, finite, static (that is, changes occur only when the planning agent acts), and discrete (in time, action, objects, and effects). Thus, the planning problem can be clearly defined and solved using a logical approach [1, 2]. Formally this is defined as a planning task = (, , ,) where:

Executing Action Languages for Planning Problems on Multi-core Platforms: Some Preliminary Results

cs.nmsu.edu

The goal of this paper is to demonstrate that parallel programming techniques can boost AI planning systems in various aspects. It shows that an appropriate parallelization of a sequential planning system often brings gain in performance and/or scalability. We start by describing general schemes for parallelizing the construction of a plan. We then discuss the applications of these techniques to two domain-independent heuristic search based planners-a competitive conformant planner (CPA) and a state-of-the-art classical planner (FF). We present experimental results which show that performance improvements and scalability are obtained in both cases. Finally, we discuss the issues that should be taken into consideration when designing a parallel planning system and relate our work to the existing literature.

Encoding plans in propositional logic

1996

In recent work we showed that planning problems can be efficiently solved by general propositional satisfiability algorithms (Kautz and Selman 1996). A key issue in this approach is the development of practical reductions of planning to SAT. We introduce a series of different SAT encodings for STRIPS-style planning, which are sound and complete representations of the original STRIPS specification, and relate our encodings to the Graphplan system of Blum and Furst (1995). We analyze the size complexity of the various encodings, both in terms of number of variables and total length of the resulting formulas. This paper complements the empirical evaluation of several of the encodings reported in Kautz and Selman (1996). We also introduce a novel encoding based on the theory of causal planning, that exploits the notion of "lifting" from the theorem-proving community. This new encoding strictly dominates the others in terms of asymptotic complexity. Finally, we consider further reductions in the number of variables used by our encodings, by compiling away either statevariables or action-variables.