Search

Intelligent agents must have a model of the dynamics of the domain in which they act. Models can be encoded by human experts or, as required by autonomous systems, automatically acquired from observation. At the state of the art, there existseveral systems for automated acquisition of planning domain models.\ud \ud \ud In this paper we present a brief overview of the automated tools that can be exploited to induce planning domain models. While reviewing the literature on the existing tools for Knowledge Engineering (KE), we do a comparative analysis of them. The analysis is based on a set of criteria.\ud \ud \ud The aim of the analysis is to give insights into the strengths and weaknesses of the considered systems, and to provide input for new, forthcoming research on KE tools in order to address future challenges in the automated KE area.

It needs much effort to encode a planning domain model for complex applications. It includes getting the domain knowledge, creating and validating the correctness of knowledge and then maintaining it. The major product is the domain model which consists of a set of operators [1] (also called action schema) written in certain modelling languages. In this paper we briefly review a variety of automated knowledge engineering tools that automatically produce domain models, and discuss the motivation\ud behind the development of these tools with three major concerns that differentiate these tools from\ud one another. The results of the review give insights into the strengths and weaknesses of the\ud considered systems, and point to needs in future work to enhance the capabilities and overcome their short comings.

In the recent years the field of automated planing has significantly advanced and several powerful domain-independent planners have been developed. However, none of these systems clearly outperforms all the others in every known benchmark domain. This observation motivated the idea of configuring and exploiting a portfolio of planners to achieve better performances than any individual planner: some recent planning systems based on this idea obtained significantly good results in experimental analysis and International Planning Competitions. Such results lead us to think that future challenges for the automated planning community will converge on designing different approaches for combining existing planning algorithms.\ud \ud This paper focuses on the challenges and open issues of existing approaches and highlights the possible future evolution of these techniques. In addition the paper introduces algorithm portfolios, reviews existing techniques, and describes the decisions that have to be taken during the configuration.

Movement of Units in Real-Time Strategy (RTS) Games is a non-trivial and challenging task mainly due to three factors which are constraints on CPU and memory usage, dynamicity of the game world, and concurrency. In this paper, we are focusing on finding a novel solution for solving the pathfinding problem in RTS Games for the units which are controlled by the computer. The novel solution combines two AI Planning approaches: Hierarchical Task Network (HTN) and Real-Time Heuristic Search (RHS). In the proposed solution, HTNs are used as a dynamic abstraction of the game map while RHS works as planning engine with interleaving of plan making and action executions. The article provides algorithmic details of the model while the empirical details of the model are obtained by using a real-time strategy game engine called ORTS (Open Real-time Strategy). The implementation of the model and its evaluation methods are in progress however the results of the automatic HTN creation are obtained for a small scale game map.

Our research concerns formal, expressive, object centred languages and tools for use in engineering domains for planning applications. In this paper we extend our recent work on an object-centred language for encoding precondition\ud planning domains to a language called OCLh,\ud designed for HTN planning. Domain encodings for HTN planners are particularly troublesome, because they tend to be used in knowledged-based applications requiring a great deal of ‘domain engineering’, and the abstract operators\ud central to an HTN model do not share the fairly clear declarative semantics of concrete pre- and post condition operators. Central to our approach is the development, in parallel, of the abstract operator set and the hierarchical state specification of the objects that the operators manipulate. In this paper we define and illustrate a transparency property, together with a transparency checking tool, which helps the developer to encode a clear planning model in\ud OCLh. Our encoding of the Translog domain is used as an extended example to illustrate the approach.

In this paper we examine the specification and validation of Artificial Intelligence Planning domain models using the B Abstract Machine Notation and its associated tool support. We compare this to the use of OCL (object-centred language) within its tool-supported environment, GIPO. We present encodings of two well-known AI planning domain models, the Blocks world and the Tyres world, with the aim of finding a correspondence between the B and the OCL languages. We also compare the tool-supported validation offered by their respective environments.

In this paper we postulate OCLh as a prototype for future planning domain modelling languages which are foundationally sound, but offer features that are attractive and supportive to knowledge engineers. The novel contributions of this paper is that it (a) describes a truth criterion for OCLh and details a proof that the criterion is sufficient for ensuring necessary truth in a partial plan structure (b) evaluates OCLh, illustrating its pragmatic benefits by comparing it with O-Plan’s TF. We show using a real example how OCLh’s structuring devices aid the knowledge engineer in building a model. Finally, the example and comparison with TF identifies further development work to advance OCLh as potential high level research language for modelling operator based planning domains.

Despite the advances made in the last decade in automated planning, no planner outperforms all the others in every known benchmark domain. This observation motivates the idea of selecting different planning algorithms for different domains. Moreover, the planners' performances are affected by the structure of the search space, which depends on the encoding of the considered domain. In many domains, the performance of a planner can be improved by exploiting additional knowledge, extracted in the form of macro-operators or entanglements.\ud In this paper we propose ASAP, an automatic Algorithm Selection Approach for Planning that: (i) for a given domain initially learns additional knowledge, in the form of macro-operators and entanglements, which is used for creating different encodings of the given planning domain and problems, and (ii) explores the 2 dimensional space of available algorithms, defined as encodings--planners couples, and then (iii) selects the most promising algorithm for optimising either the runtimes or the quality of the solution plans.

The validation of planning domain models is an important issue and can present problems. In this paper we describe ongoing work which attempts to overcome these problems through the construction of a B-AMN specication which models the domain. The B-Method utilises B-AMN, a state-based formal specication language with tool support provided by the B-Toolkit. We describe how this tool support provides facilities for both animation and proof, and we propose the use of the B-Method at an initial level of domain capture. The approach is illustrated using a simple transport domain, where a high-level model, which can be reasoned with, is produced. In addition, animation allows validation by users.

Domain-independent planning systems require that domain\ud constraints and invariants are specified as part of\ud the input domain model. In AI Planning, the generated\ud plan is correct provided the constraints of the world\ud in which the agent is operating are satisfied. Specifying\ud operator descriptions by hand for planning domain\ud models that also require domain specific constraints is\ud time consuming, error prone and still a challenge for the\ud AI planning community.\ud The LOCM (Cresswell, McCluskey, and West 2013)\ud system carries out automated generation of the dynamic\ud aspects of a planning domain model from a set of example\ud training plans. We enhance the output domain\ud model of the LOCM system to capture static domain\ud constraints from the same set of input training plans as\ud used by LOCM to learn dynamic aspects of the world.\ud In this paper we propose a new framework ASCoL (Automated\ud Static Constraint Learner), to make constraint\ud acquisition more efficient, by observing a set of training\ud plan traces. Most systems that learn constraints automatically\ud do so by analysing the operators of the planning\ud world. Out proposed system will discover static\ud constraints by analysing plan traces for correlations in\ud the data. To do this an algorithm is in the process of\ud development for graph discovery from the collection of\ud ground action instances used in the input plan traces.\ud The proposed algorithm will analyse the complete set of\ud plan traces, based on a predefined set of constraints, and\ud deduces facts from it. We then augment components of\ud the LOCM generated domain with enriched constraints.

The core task of automated planning is goal-directed action selection; this task is not unique to the planning community, but is also relevant to numerous other research areas within AI. One such area is interactive systems, where a fundamental component called the interaction manager selects actions in the context of conversing with humans using natural language. Although this has obvious parallels to automated planning, using a planner to address the interaction management task relies on appropriate engineering of the underlying planning domain and planning problem to capture the necessary dynamics of the world, the agents involved, their actions, and their knowledge. In this chapter, we describe work on using domain-independent automated planning for action section in social human-robot interaction, focusing on work from the JAMES (Joint Action for\ud Multimodal Embodied Social Systems) robot bartender project.