Probabilistic independence has proved to be a fundamental tool that can dramatically simplify the task of eliciting, representing, and computing with probabilities. We advance the position that notions of utility independence can serve similar functions when reasoning about preferences and utilities during decision making.

this paper we introduce Prohairetic Deontic Logic (PDL), a preference-based

this article we analyze this overlap, and we also show that this confusion can be avoided if one makes the proper distinctions between different types of defeasibility. Furthermore, we also show that these distinctions are essential for an adequate analysis of notorious contrary-to-duty paradoxes such as the Chisholm and Forrester paradoxes. The main claim of this article is that the defeasible aspect of defeasible deontic logic is different from the defeasible aspect of, for example, Reiter's default rules (Reiter, 1980). Different types of defeasibility in a logic of defeasible reasoning formalize a single notion, whereas defeasible deontic logics formalize two notions. Consider first the logics of defeasible reasoning and the famous Tweety example. In the case of factual defeasibility, we say that the `birds fly' default is cancelled by the fact :f , and in the case of overridden defeasibility by the `penguins do not fly' default. By cancellation we mean, for example, that if :f is true, then the default assumption that f

Autonomous agents reason frequently about preferences such as desires and goals. In this paper we propose a logic of desires with a utilitarian semantics, in which we study non-monotonic reasoning about desires and preferences based on the idea that desires can be understood in terms of utility losses (penalties for violations) and utility gains (rewards for fulllments). Our logic allows for a systematic study and classication of desires, for example by distinguishing subtly dierent ways to add up these utility losses and gains. We propose an explicit construction of the agent's preference relation from a set of desires together with dierent kinds of knowledge. A set of desires extended with knowledge induces a set of `distinguished' utility functions by adding up the utility losses and gains of the individual desires, and these distinguished utility functions induce the preference relation.

Logics of desires, preferences and goals have recently been proposed in planning and agent theory. In this paper we introduce a dynamic logic with utilitarian desires and we discuss the relation between desires and utilities. Given that an agent's desires refer to his utility function, which we assume to be constant, we resolve the paradox that the stability of the agent's utility function does not imply the stability of his desires. We illustrate the use of the logic to formalize certain aspects of negotiation. In particular, we show how one agent can inuence the behavior of another agent by inuencing his desires.

In this paper we introduce Prohairetic Deontic Logic, a preference-based dyadic deontic logic. An obligation `ff should be (done) if fi is (done)' is true if (1) no :ff fi is preferred to an fffi and (2) the preferred fi are ff. We show that this mixed representation solves several problems of deontic logic. Moreover, we discuss the relation between preferencebased deontic logic and qualitative decision theory. Introduction Deontic logic is a modal logic developed in philosophical logic. Absolute and conditional obligations are represented by the modal formulas Off and O(ff j fi), where the latter is read as `ff ought to be (done) if fi is (done).' The development of preference-based deontic logic is plagued by the following three (related) problems: Strong preference problem. Preferences for ff 1 and ff 2 conflict for ff 1 :ff 2 and :ff 1 ff 2 . Contrary-to-duty problem. A contrary-to-duty obligation is an obligation that refers to a sub-ideal situation. Reasoning structures lik...

... this paper we argue that the inclusion of both desires and goals in the same model can be important, particularly in a Multi-Agent System context, where other sources of individual motivation such as obligations and norms may be present. This leads us to propose an extended BDI architecture in which obligations, norms and desires are distinguished from goals and explicitly represented. In this paper we consider suitable logical representations for and properties of these elements, and describe the basic method of operation of the architecture, focusing on how goal generation and goal maintenance may occur.

. In this paper we study the logic of goals, which are formalized as desires with an utilitarian semantics. In our framework goals have a dual character, because they are constraints on utility functions as well as constructors of these utility functions. The non-monotonic reasoning related to the constructors reflects that goals are used as heuristic approximations of preferences in decision making and planning. Moreover, our framework is based on bipolar additive preferences, where bipolarity means that goals can either result in a gain of utility if achieved, or a loss of utility if not achieved. The framework is used to illustrate different types of context-dependence and conflicts of goals. 1 Introduction Decision theory has become widely accepted in the AI community as a useful framework for planning and decision making [9]. In the context of qualitative decision theory [15, 18, 2] recently several logics for goals and desires have been proposed [8, 7, 3, 15, 17, 16, 14, 12, 20...

A theory of diagnosis and qualitative decision theory are able to formalize reasoning with norms. They are thus different from deontic logic, that formalizes reasoning about norms. In this paper, we compare two theories of diagnosis for normative systems: Ramos and Fiadeiro's theory of diagnosis developed for organizational process design and Tan and Van der Torre's theory of diagnosis extended with notions of qualitative decision theory. We observe several similarities. 1 Introduction In this paper we argue that normative reasoning is more than deontic logic. Deontic logic tells you which obligations can be derived from a set of other obligations. In particular, it characterizes the logical relations between obligations. For example, in most deontic logics the conjunction p q is obliged, if both p and q are obliged. However, it does not explain how norms effect the behavior of rational agents. From Op you cannot infer whether somebody will actually perform p. This is no critique on ...

In this paper we introduce the diagnostic and decision- theoretic framework for deontic reasoning diO(de) 2 . The framework diO(de) 2 formalizes reasoning about violations and goals. We use this framework to discuss two theories of normative reasoning, diagnosis theory and (qualitative) decision theory. A crucial distinction between the two theories is their perspective on time. Diagnosis theory reasons about incomplete knowledge and only considers the past. It distinguishes between violations and non-violations. Qualitative decision theory reasons about decision variables and considers the future. It distinguishes between fulfilled obligations and unfulfilled obligations. Moreover, we discuss the relation between the two theories of normative reasoning and deontic logic. The theories formalize reasoning with norms, and they are thus different from deontic logic, that formalizes reasoning about norms. 1 Introduction There is a discussion in AI and law literature whether deontic...