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...

We pursue the perspective of Reiter that in the situation calculus one can formalize primitive, determinate actions with axioms which, among others, include two disjoint sets: a set of successor state axioms and a set of action precondition axioms. We posed ourselves the problem of automatically generating successor state axioms, given only a set of effect axioms and a set of state constraints. This is a special version of what has been traditionally called the ramification problem. To our surprise, we found that there are state constraints whose role is not to yield indirect effects of actions. Rather, they are implicit axioms about action preconditions. As such, they are intimately related to the classical qualification problem. We also discovered that other kinds of state constraints arise; these are related to the formalization of strategic or control information. This paper is devoted to describing our results along these lines, focusing on ramification and qualification state con...

This paper considers the problem of specifying the effects of actions in the situation calculus using domain constraints. We argue that normal state constraints that refer to only the truth values of fluents are not strong enough for this purpose, and that a notion of causation needs to be employed explicitly. Technically, we introduce a new ternary predicate Caused into the situation calculus: Caused(p; v; s) if the proposition p is caused (by something unspecified) to have the truth value v in the state s. Using this predicate, we can represent not only action-triggered causal statements such as that the action load causes the gun to be loaded, but also fluent-triggered ones such as that the fact that the switch is in the up position causes the lamp to be on. The former is convenient for representing the direct effects of actions, and the latter the indirect effects. 1 Introduction We consider the problem of formalizing the effects of actions in the situation calculus [ McCarthy a...

The ramification problem in the context of commonsense reasoning about actions and change names the challenge to accommodate actions whose execution causes indirect effects. Not being part of the respective action specification, such effects are consequences of general laws describing dependencies between components of the world description. We present a general approach to this problem which incorporates causality, formalized by directed relations between two single effects stating that, under specific circumstances, the occurrence of the first causes the second. Moreover, necessity of exploiting causal information in this way or a similar is argued by elaborating the limitations of common paradigms employed to handle ramifications, namely, the principle of categorization and the policy of minimal change. Our abstract solution is exemplarily integrated into a specific calculus based on the logic programming paradigm. To apper in: Artificial Intelligence Journal On leave from FG Inte...

Abduction is an important form of nonmonotonic reasoning allowing one to find explanations for certain symptoms or manifestations. When the application domain is described by a logical theory, we speak about logic-based abduction. Candidates for abductive explanations are usually subjected to minimality criteria such as subsetminimality, minimal cardinality, minimal weight, or minimality under prioritization of individual hypotheses. This paper presents a comprehensive complexity analysis of relevant decision and search problems related to abduction on propositional theories. Our results indicate that abduction is harder than deduction. In particular, we show that with the most basic forms of abduction the relevant decision problems are complete for complexity classes at the second level of the polynomial hierarchy, while the use of prioritization raises the complexity to the third level in certain cases.

Reasoning about change is an important aspect of commonsense reasoning and planning. In this paper we describe an approach to reasoning about change for rich domains where it is not possible to anticipate all situations that might occur. The approach provides a solution to the frame problem, and to the related problem that it is not always reasonable to explicitly specify all of the consequences of actions. The approach involves keeping a single model of the world that is updated when actions are performed. The update procedure involves constructing the nearest world to the current one in which the consequences of the actions under consideration hold. The way we find the nearest world is to construct proofs of the negation of the explicit consequences of the expected action, and to remove a premise in each proof from the current world. Computationally, this construction procedure appears to be tractable for worlds like our own where few things tend to change with each action, or where ...

In the situation calculus, it is sometimes necessary to prove that certain properties are true in all world states accessible from the initial state. This is the case for some forms of reasoning about the physical world, for certain planning applications, and for verifying integrity constraints in databases. Not surprisingly, this requires a suitable form of mathematical induction. This paper motivates the need for proving properties of states in the situation calculus, proposes appropriate induction principles for this task, and gives examples of their use in databases and for reasoning about the physical world. Abbreviated title: Proving Properties of States 1 Introduction The situation calculus [8] is enjoying new popularity these days. One reason is that its expressiveness is considerably richer than has been commonly believed (Gelfond, Lifschitz and Rabinov [2], Pinto and Reiter [10], Schubert [16]). Another is the possibility of precisely characterizing the strengths and limi...

this paper, including transaction logs and historical queries, the complexity of query evaluation, actualized transactions, logic programming approaches to updates, database views and state constraints. / This paper consolidates and expands on a variety of results, some of which have been described elsewhere (Reiter [46, 45, 44])

We present a computationally effective approach to representing and reasoning about actions with many qualifications. The approach involves treating actions as qualified not by specific facts that may or may not hold when the action is executed, but instead as potentially qualified by general constraints describing the domain being investigated. Specifically, we suggest that the result of the action be computed without considering these qualifying domain constraints, and take the action to be qualified if and only if any of the constraints is violated after the computation is complete. Our approach is presented using the framework developed in [6], where we discussed a solution to the frame and ramification problems based on the notion of possible worlds, and compared the computational requirements of that solution to the needs of more conventional ones. In the present paper, we show that the domain constraint approach to qualification, coupled with the possible worlds approach describ...

We investigate the problem of reasoning with partitions of related logical axioms. Our motivation is two-fold. First, we are concerned with how to reason effectively with multiple knowledge bases that have overlap in content. Second, and more fundamentally, we are concerned with how to exploit structure inherent in a set of logical axioms to induce a partitioning of the axioms that will lead to an improvement in the efficiency of reasoning. To this end, we provide algorithms for reasoning with partitions of axioms in propositional and first-order logic. Craig’s interpolation theorem serves as a key to proving completeness of these algorithms. We analyze the computational benefit of our algorithms and detect those parameters of a partitioning that influence the efficiency of computation. These parameters are the number of symbols shared by a pair of partitions, the size of each partition, and the topology of the partitioning. Finally, we provide a greedy algorithm that automatically decomposes a given theory into partitions, exploiting the parameters that influence the efficiency of computation. 1