The concept of an agent has become important in both Artificial Intelligence (AI) and mainstream computer science. Our aim in this paper is to point the reader at what we perceive to be the most important theoretical and practical issues associated with the design and construction of intelligent agents. For convenience, we divide these issues into three areas (though as the reader will see, the divisions are at times somewhat arbitrary). Agent theory is concerned with the question of what an agent is, and the use of mathematical formalisms for representing and reasoning about the properties of agents. Agent architectures can be thought of as software engineering models of agents; researchers in this area are primarily concerned with the problem of designing software or hardware systems that will satisfy the prop-erties specified by agent theorists. Finally, agent languages are software systems for programming and experimenting with agents; these languages may embody principles proposed by theorists. The paper is not intended to serve as a tutorial introduction to all the issues mentioned; we hope instead simply to identify the most important issues, and point to work that elaborates on them. The article includes a short review of current and potential applications of agent technology.

This paper presents a simple, sound, complete, and systematic algorithm for domain independent STRIPS planning. Simplicity is achieved by starting with a ground procedure and then applying a general, and independently verifiable, lifting transformation. Previous planners have been designed directly as lifted procedures. Our ground procedure is a ground version of Tate's NONLIN procedure. In Tate's procedure one is not required to determine whether a prerequisite of a step in an unfinished plan is guaranteed to hold in all linearizations. This allows Tate's procedure to avoid the use of Chapman's modal truth criterion. Systematicity is the property that the same plan, or partial plan, is never examined more than once. Systematicity is achieved through a simple modification of Tate's procedure.

Planning under uncertainty is a central problem in the study of automated sequential decision making, and has been addressed by researchers in many different fields, including AI planning, decision analysis, operations research, control theory and economics. While the assumptions and perspectives adopted in these areas often differ in substantial ways, many planning problems of interest to researchers in these fields can be modeled as Markov decision processes (MDPs) and analyzed using the techniques of decision theory. This paper presents an overview and synthesis of MDP-related methods, showing how they provide a unifying framework for modeling many classes of planning problems studied in AI. It also describes structural properties of MDPs that, when exhibited by particular classes of problems, can be exploited in the construction of optimal or approximately optimal policies or plans. Planning problems commonly possess structure in the reward and value functions used to de...

Learning, planning, and representing knowledge at multiple levels of temporal abstraction are key, longstanding challenges for AI. In this paper we consider how these challenges can be addressed within the mathematical framework of reinforcement learning and Markov decision processes (MDPs). We extend the usual notion of action in this framework to include options---closed-loop policies for taking action over a period of time. Examples of options include picking up an object, going to lunch, and traveling to a distant city, as well as primitive actions such as muscle twitches and joint torques. Overall, we show that options enable temporally abstract knowledge and action to be included in the reinforcement learning framework in a natural and general way. In particular, we show that options may be used interchangeably with primitive actions in planning methods such as dynamic programming and in learning methods such as Q-learning.

This paper presents a new approach to hierarchical reinforcement learning based on decomposing the target Markov decision process (MDP) into a hierarchy of smaller MDPs and decomposing the value function of the target MDP into an additive combination of the value functions of the smaller MDPs. The decomposition, known as the MAXQ decomposition, has both a procedural semantics---as a subroutine hierarchy---and a declarative semantics---as a representation of the value function of a hierarchical policy. MAXQ unifies and extends previous work on hierarchical reinforcement learning by Singh, Kaelbling, and Dayan and Hinton. It is based on the assumption that the programmer can identify useful subgoals and define subtasks that achieve these subgoals. By defining such subgoals, the programmer constrains the set of policies that need to be considered during reinforcement learning. The MAXQ value function decomposition can represent the value function of any policy that is consisten...

O-Plan is an AI planner based on previous experience with the Nonlin planner and its derivatives. Nonlin and other similar planning systems had limited control architectures and were only partially successful at limiting their search spaces. O-Plan is a design and implementation of a more flexible system aimed at supporting planning research and development, opening up new planning methods and supporting strong search control heuristics. O-Plan takes an engineering approach to the construction of an efficient domain independent planning system which includes a mixture of AI and numerical techniques from Operations Research. The main contributions of the work are centred around the control of search within the OPlan planning framework, and this paper outlines the search control heuristics employed within the planner. These involve the use of condition typing, time and resource constraints and domain constraints to allow knowledge about an application domain to be used to prune the searc...

The concept of an agent has recently become important in Artificial Intelligence (AI), and its relatively youthful subfield, Distributed AI (DAI). Our aim in this paper is to point the reader at what we perceive to be the most important theoretical and practical issues associated with the design and construction of intelligent agents. For convenience, we divide the area into three themes (though as the reader will see, these divisions are at times somewhat arbitrary). Agent theory is concerned with the question of what an agent is, and the use of mathematical formalisms for representing and reasoning about the properties of agents. Agent architectures can be thought of as software engineering models of agents; researchers in this area are primarily concerned with the problem of constructing software or hardware systems that will satisfy the properties specified by agent theorists. Finally, agent languages are software systems for programming and experimenting with agents; these languages typically embody principles proposed by theorists. The paper is not intended to serve as a tutorial introduction to all the issues mentioned; we hope instead simply to identify the key issues, and point to work that elaborates on them. The paper closes with a detailed bibliography, and some bibliographical remarks. 1

Over the years increasingly sophisticated planning algorithms have been developed. These have made for more efficient planners, but unfortunately these planners still suffer from combinatorial complexity even in simple domains. Theoretical results demonstrate that planning is in the worst case intractable. Nevertheless, planning in particular domains can often be made tractable by utilizing additional domain structure. In fact, it has long been acknowledged that domain independent planners need domain dependent information to help them plan effectively. In this

This article presents a completely automated approach to generating abstractions for planning. The abstractions are generated using a tractable, domain-independent algorithm whose only input is the definition of a problem to be solved and whose output is an abstraction hierarchy that is tailored to the particular problem. The algorithm generates abstraction hierarchies by dropping literals from the original problem definition. It forms abstractions that satisfy the ordered monotonicity property, which guarantees that the structure of an abstract solution is not changed in the process of refining it. The algorithm for generating abstractions is implemented in a system called alpine, which generates abstractions for a hierarchical version of the prodigy problem solver. The abstractions generated by alpine are tested in multiple domains on large problem sets and are shown to produce shorter solutions with significantly less search than planning without using abstraction. 1 1 ...

Computational efficiency is a central concern in the design of knowledge representation systems. In order to obtain efficient systems, it has been suggested that one should limit the form of the statements in the knowledge base or use an incomplete inference mechanism. The former approach is often too restrictive for practical applications, whereas the latter leads to uncertainty about exactly what can and cannot be inferred from the knowledge base. We present a third alternative, in which knowledge given in a general representation language is translated (compiled) into a tractable form — allowing for efficient subsequent query answering. We show how propositional logical theories can be compiled into Horn theories that approximate the original information. The approximations bound the original theory from below and above in terms of logical strength. The procedures are extended to other tractable languages (for example, binary clauses) and to the first-order case. Finally, we demonstrate the generality of our approach by compiling concept descriptions in a general framebased language into a tractable form.