We present some techniques for planning in domains specified with the recent standard language pddl2.1, supporting “durative actions ” and numerical quantities. These tech-niques are implemented in lpg, a domain-independent planner that took part in the 3rd International Planning Competition (IPC). lpg is an incremental, any time system pro-ducing multi-criteria quality plans. The core of the system is based on a stochastic local search method and on a graph-based representation called “Temporal Action Graphs ” (TA-graphs). This paper focuses on temporal planning, introducing TA-graphs and proposing some techniques to guide the search in lpg using this representation. The experimental results of the 3rd IPC, as well as further results presented in this paper, show that our techniques can be very effective. Often lpg outperforms all other fully-automated plan-ners of the 3rd IPC in terms of speed to derive a solution, or quality of the solutions that can be produced. 1.

During the last five years, the planning community has seen vast progress in terms of the sizes of benchmark examples that domain-independent planners can tackle successfully. The key technique behind this progress is the use of heuristic functions based on relaxing the planning task at hand, where the relaxation is to assume that all delete lists are empty. The success of such methods in many of the current benchmarks suggests that in those task's state spaces relaxed goal distances yield a heuristic function of high quality.

Heuristic search methods promise to find shortest paths for path-planning problems faster than uninformed search methods. Incremental search methods, on the other hand, promise to find shortest paths for series of similar path-planning problems faster than is possible by solving each path-planning problem from scratch. In this article, we develop Lifelong Planning A * (LPA*), an incremental version of A * that combines ideas from the artificial intelligence and the algorithms literature. It repeatedly finds shortest paths from a given start vertex to a given goal vertex while the edge costs of a graph change or vertices are added or deleted. Its first search is the same as that of a version of A * that breaks ties in favor of vertices with smaller g-values but many of the subsequent searches are potentially faster because it reuses those parts of the previous search tree that are identical to the new one. We present analytical results that demonstrate its similarity to A * and experimental results that demonstrate its potential advantage in two different domains if the path-planning problems change only slightly and the changes are close to the goal.

Many real-world planning problems require one to solve a series of similar planning tasks. In this case, replanning can be much faster than planning from scratch. In this paper, we introduce a novel replanning method for symbolic planning with heuristic search-based planners, currently the most popular planners. Our SHERPA replanner is not only the first heuristic search-based replanner but, different from previous replanners for other planning paradigms, it also guarantees that the quality of its plans is as good as that achieved by planning from scratch. We provide an experimental feasibility study that demonstrates the promise of SHERPA for heuristic search-based replanning.

Abstract—This paper deals with building up a knowledge base of manipulation tasks by extracting relevant knowledge from demonstrations of manipulation problems. Hereby the focus of the paper is on modeling and representing manip-ulation tasks enabling the system to reason and reorganize the gathered knowledge in terms of reusability, scalability and explainability of learned skills and tasks. The goal is to compare the newly acquired skill or task with already existing task knowledge and decide whether to add a new task rep-resentation or to expand the existing representation with an alternative. Furthermore, a constraint for the representation is that at execution time the built knowledge base can be integrated and used in a symbolic planner. Index Terms—Modeling of Manipulation Tasks, Program-ming by Demonstration, Reasoning about Tasks and Skills

www.cs.yorku.ca/~wolfgang/ Fractal terrains provide an easy way to generate realistic landscapes. There are several methods to generate fractal terrains, but none of those algorithms allow the user much flexibility in controlling the shape or properties of the final outcome. A few methods to modify fractal terrains have been previously proposed, both algorithm-based as well as by hand editing, but none of these provide a general solution. In this work, we present a new algorithm for fractal terrain deformation. We present a general solution that can be applied to a wide variety of deformations. Our approach employs stochastic local search to identify a sequence of local modifications, which deform the fractal terrain to conform to a set of specified constraints. The presented results show that the new method can incorporate multiple constraints simultaneously, while still preserving the natural look of the fractal terrain. Keywords: (according to ACM CCS): I.3.7 [Computer Graphics, Three-Dimensional Graphics and Realism]: Fractals, I.2.8 [Problem Solving, Control Methods, and Search] Graph and tree search strategies

Planning is often not a one-shot task because either the world or the agent’s knowledge of the world changes. In this paper, we introduce a new principle that can be used to solve a series of similar search tasks faster with heuristic search methods than running individ-ual searches in isolation, by updating the heuristics over time to make them more informed and thus fu-ture searches more focused. This principle is simple and easy to integrate into heuristic search methods, and it is easy to prove the correctness of the resulting heuristic search methods.

In this paper, we will describe a value-driven graph search technique that is capable of generating a rich variety of single and multiple vehicle behaviors. The generation of behaviors depends on cost and benefit computations that may involve terrain characteristics, line of sight to enemy positions, and cost, benefit, and risk of traveling on roads. Depending on mission priorities and cost values, real-time planners can autonomously build appropriate behaviors on the fly that include road following, cross-country movement, stealthily movement, formation keeping, and bounding overwatch. This system follows NIST's 4D/RCS architecture, and a discussion of the world model, value judgment, and behavior generation components is provided. In addition, techniques for collapsing a multidimensional model space into a cost space and planning graph constraints are discussed. The work described in this paper has been performed under the Army Research Laboratory's Robotics Demo III program.

Multi-modal transportation is a logistics problem in which a set of goods have to be transported to different places, with the combination of at least two modes of transport, without a change of container for the goods. The goal of this paper is to describe TIMIPLAN, a system that solves multi-modal transportation problems in the context of a project for a big company. In this paper, we combine Linear Programming (LP) with automated planning techniques in order to obtain good quality solutions. The direct use of classical LP tech-niques is difficult in this domain, because of the non-linearity of the optimization function and constraints; and planning al-gorithms cannot deal with the entire problem due to the large number of resources involved. We propose a new hybrid algo-rithm, combining LP and planning to tackle the multi-modal transportation problem, exploiting the benefits of both kinds of techniques. The system also integrates an execution com-ponent that monitors the execution, keeping track of failures and replans if necessary, maintaining most of the plan in exe-cution. We also present some experimental results that show the performance of the system.

Intermodal transportation is a very challenging task for AI planning. It is a logistics problem in which a set of goods have to be transported to different places, with the combina-tion of at least two modes of transport in a single transport chain, without a change of container for the goods. The main goal of this paper is to introduce TIMIPLAN, a new applica-tion to solve multimodal transportation problems. TIMIPLAN has been developed in the context of a research project in-volving one of the biggest spanish companies in intermodal transportation, Acciona Transmediterránea Cargo. The main challenge of this project is the size of the planning problems: more than 300 containers, trucks, locations and services have to be dealt with everyday. Thus, internally, TIMIPLAN com-bines Operational Research (OR) techniques with Artificial Intelligence (AI) planning in order to obtain good quality plans, by exploiting the benefits of both kinds of techniques. These plans can be also graphically visualized.