Utility Functions for Ceteris Paribus Preferences

Ceteris paribus (all-else equal) preference statements concisely represent preferences over outcomes or goals in a way natural to human thinking. Although deduction in a logic of such statements can compare the desirability of specific conditions or goals, many decision-making methods require numerical measures of degrees of desirability. To permit ceteris paribus specifications of preferences while providing quantitative comparisons, we present an algorithm that compiles a set of qualitative ceteris paribus preferences into an ordinal utility function. Our algorithm is complete for a finite universe of binary features. Constructing the utility function can, in the worst case, take time exponential in the number of features, but common independence conditions reduce the computational burden. We present heuristics using utility independence and constraint-based search to obtain efficient utility functions.     

Reasoning about Actions and Planning with Preferences Using Prioritized Default Theory

This paper shows how action theories, expressed in an extended version of the language     , can be naturally encoded using Prioritized Default Theory . We also show how prioritized default theory can be extended to express preferences between rules . This extension provides a natural framework to introduce different types of preferences in action theories— preferences between actions and preferences between final states . In particular, we demonstrate how these preferences can be expressed within extended prioritized default theory. We also discuss how this framework can be implemented in terms of answer set programming.     

Using arguments for making and explaining decisions

Arguments play two different roles in day life decisions, as well as in the discussion of more crucial issues. Namely, they help to select one or several alternatives, or to explain and justify an already adopted choice.This paper proposes the first general and abstract argument-based framework for decision making. This framework follows two main steps. At the first step, arguments for beliefs and arguments for options are built and evaluated using classical acceptability semantics. At the second step, pairs of options are compared using decision principles. Decision principles are based on the accepted arguments supporting the options. Three classes of decision principles are distinguished: unipolar, bipolar or non-polar principles depending on whether i) only arguments pros or only arguments cons, or ii) both types, or iii) an aggregation of them into a meta-argument are used. The abstract model is then instantiated by expressing formally the mental states (beliefs and preferences) of a decision maker. In the proposed framework, information is given in the form of a stratified set of beliefs. The bipolar nature of preferences is emphasized by making an explicit distinction between prioritized goals to be pursued, and prioritized rejections that are stumbling blocks to be avoided. A typology that identifies four types of argument is proposed. Indeed, each decision is supported by arguments emphasizing its positive consequences in terms of goals certainly satisfied and rejections certainly avoided. A decision can also be attacked by arguments emphasizing its negative consequences in terms of certainly missed goals, or rejections certainly led to by that decision. Finally, this paper articulates the optimistic and pessimistic decision criteria defined in qualitative decision making under uncertainty, in terms of an argumentation process. Similarly, different decision principles identified in multiple criteria decision making are restated in our argumentation-based framework.     

On the revision of preferences and rational inference processes

Orderings and inference relations can be successfully used to model the behavior of a rational agent. This behavior is indeed represented either by a set of ordered pairs that reflect the agent's preferences, or by a rational inference relation that describes the agent's internal logics. In the finite case where we work, both structures admit a simple representation by means of logical chains. The problem of revising such inference processes arises when it appears necessary to modify the original model in order to take into account new facts about the agent's behavior. How is it then possible to perform the desired modification? We study here the possibilities offered by the technique of ‘chain revision’ which appears to be the easiest way to treat this kind of problem: the revision is performed through a simple modification of the logical chain attached to the agent's behavior, and the revision problem boils down to adding, retracting or modifying some of the links of the original chain. This perspective permits an effective treatment of the problems of both simple and multiple revision. The technique developed can also be used in some limiting cases, when the agent's inference process is only partially known, encoded by an incomplete set of preferences or a conditional knowledge base.     

Reasoning with Sets of Defaults in Default Logic

We present a general approach for representing and reasoning with sets of defaults in default logic, focusing on reasoning about preferences among sets of defaults. First, we consider how to control the application of a set of defaults so that either all apply (if possible) or none do (if not). From this, an approach to dealing with preferences among sets of default rules is developed. We begin with an ordered default theory , consisting of a standard default theory, but with possible preferences on sets of rules. This theory is transformed into a second, standard default theory wherein the preferences are respected. The approach differs from other work, in that we obtain standard default theories and do not rely on prioritized versions of default logic. In practical terms this means we can immediately use existing default logic theorem provers for an implementation. Also, we directly generate just those extensions containing the most preferred applied rules; in contrast, most previous approaches generate all extensions, then select the most preferred. In a major application of the approach, we show how semimonotonic default theories can be encoded so that reasoning can be carried out at the object level. With this, we can reason about default extensions from within the framework of a standard default logic. Hence one can encode notions such as skeptical and credulous conclusions, and can reason about such conclusions within a single extension.     

Some realizations and instances of Yager prioritized preference frame with application in evaluation and decision making

This study first revamps Yager prioritized ordered weighted averaging operators, and condenses them into a conceptual frame with putting aside one realization from Yager's original proposal. Then, based on elicited conceptual frame called Yager prioritized preference conceptual frame, this article proposes three distinct realizations to the conceptual frame with corresponding different instances, in which some evaluation models with weights determination methods are provided. Numerical examples are also presented immediately after every instance. Lastly, this study proposes the concepts of outer monotonic, inner monotonic, and global monotonic weights functions, and discusses some related properties, which are often embodied in preferences of decision makers.     

Utilitarian Desires

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 nonmonotonic 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 fulfillments). Our logic allows for a systematic study and classification of desires, for example by distinguishing subtly different 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 different 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.     

Integrating Multiple Learning Strategies in First Order Logics

This paper describes a representation framework that offers a unifying platform for alternative systems, which learn concepts in First Order Logics. The main aspects of this framework are discussed. First of all, the separation between the hypothesis logical language (a version of the VL21 language) and the representation of data by means of a relational database is motivated. Then, the functional layer between data and hypotheses, which makes the data accessible by the logical level through a set of abstract properties is described. A novelty, in the hypothesis representation language, is the introduction of the construct of internal disjunction; such a construct, first used by the AQ and Induce systems, is here made operational via a set of algorithms, capable to learn it, for both the discrete and the continuous-valued attributes case. These algorithms are embedded in learning systems (SMART+, REGAL, SNAP, WHY, RTL) using different paradigms (symbolic, genetic or connectionist), thus realizing an effective integration among them; in fact, categorical and numerical attributes can be handled in a uniform way. In order to exemplify the effectiveness of the representation framework and of the multistrategy integration, the results obtained by the above systems in some application domains are summarized.     

Assumption‐Based Argumentation Equipped with Preferences and its Application to Decision Making,Practical Reasoning,and Epistemic Reasoning

The existing approaches that map the given explicit preferences into standard assumption‐based argumentation (ABA) frameworks reveal some difficulties such as generating a huge number of rules. To overcome them, we present an assumption‐based argumentation framework equipped with preferences (p_ABA). It increases the expressive power of ABA by incorporating preferences between sentences into the framework. The semantics of p_ABA is given by extensions, which are maximal among extensions of ABA with regard to the extension ordering “lifted” from the given sentence ordering. As a theoretical contribution of this study, we show that prioritized logic programming can be formulated as a specific form of p_ABA. The advantage of our approach is that not only does p_ABA enable us to express different kinds of preferences such as preferences over rules, over goals, or over decisions by means of sentence orderings but we can also successfully obtain solutions from extensions of the p_ABA expressing the respective knowledge for various applications such as epistemic reasoning, practical reasoning, and decision making with preferences in a uniform and domain‐independent way without suffering from difficulties of the existing approaches.     

Utility Models for Goal‐Directed, Decision‐Theoretic Planners

AI planning agents are goal-directed : success is measured in terms of whether an input goal is satisfied. The goal gives structure to the planning problem, and planning representations and algorithms have been designed to exploit that structure. Strict goal satisfaction may be an unacceptably restrictive measure of good behavior, however.
A general decision-theoretic agent, on the other hand, has no explicit goals: success is measured in terms of an arbitrary preference model or utility function defined over plan outcomes. Although it is a very general and powerful model of problem solving, decision-theoretic choice lacks structure, which can make it difficult to develop effective plan‐generation algorithms.
This paper establishes a middle ground between the two models. We extend the traditional AI goal model in several directions: allowing goals with temporal extent, expressing preferences over partial satisfaction of goals, and balancing goal satisfaction against the cost of the resources consumed in service of the goals. In doing so we provide a utility model for a goal-directed agent.
An important quality of the proposed model is its tractability. We claim that our model, like classical goal models, makes problem structure explicit. This structure can then be exploited by a problem-solving algorithm. We support this claim by reporting on two implemented planning systems that adopt and exploit our model.     

Constraint-based optimization and utility elicitation using the minimax decision criterion

In many situations, a set of hard constraints encodes the feasible configurations of some system or product over which multiple users have distinct preferences. However, making suitable decisions requires that the preferences of a specific user for different configurations be articulated or elicited, something generally acknowledged to be onerous. We address two problems associated with preference elicitation: computing a best feasible solution when the user's utilities are imprecisely specified; and developing useful elicitation procedures that reduce utility uncertainty, with minimal user interaction, to a point where (approximately) optimal decisions can be made. Our main contributions are threefold. First, we propose the use of minimax regret as a suitable decision criterion for decision making in the presence of such utility function uncertainty. Second, we devise several different procedures, all relying on mixed integer linear programs, that can be used to compute minimax regret and regret-optimizing solutions effectively. In particular, our methods exploit generalized additive structure in a user's utility function to ensure tractable computation. Third, we propose various elicitation methods that can be used to refine utility uncertainty in such a way as to quickly (i.e., with as few questions as possible) reduce minimax regret. Empirical study suggests that several of these methods are quite successful in minimizing the number of user queries, while remaining computationally practical so as to admit real-time user interaction.     

Standard goal programming with fuzzy hierarchies: a sequential approach

This paper proposes a pragmatic model for multi-objective decision-making processes involving clusters of objectives which have a decisional meaning for the decision maker (DM). We provide the DMs with a comfortable tool that allows them to express their preferences both by comparing criteria of the same cluster and via the comparison between the different clusters. In standard goal programming the importance of the goals is modeled by the introduction of preferential weights or/and the incorporation of pre-emptive priorities. However, in many cases the DM is not able to establish a precise preference structure. Even in the case of precise weights the solution does not match necessarily the relative weights or, in the case of precise pre-emptive priority, the result could be very restrictive. In order to overcome these drawbacks, in this paper the normalized unwanted deviations are interpreted in terms of achievement degrees of the goals and fuzzy relations are used to model the relative importance of the goals. Thus, we show how several methodologies from the fuzzy goal programming literature can be tailored for solving standard GP problems. We apply this new modeling to problems where there is a “natural” clustering between goals of the same class. We address this situation by solving two phases; in the first one each class is handled separately taking into account the hierarchy of their goals and, in the second phase, we integrate the results of the first phase and the imprecise hierarchy of the different classes. We formulate a new goal programming model called as sequential goal programming with fuzzy hierarchy model. Because many real situations involve decision making in this environment, our proposal can be a useful tool of broad application. A numerical example illustrates the methodology.     

An Evolutionary Algorithm-Based Approach to Automated Design of Analog and RF Circuits Using Adaptive Normalized Cost Functions

Typical analog and radio frequency (RF) circuit sizing optimization problems are computationally hard and require the handling of several conflicting cost criteria. Many researchers have used sequential stochastic refinement methods to solve them, where the different cost criteria can either be combined into a single-objective function to find a unique solution, or they can be handled by multiobjective optimization methods to produce tradeoff solutions on the Pareto front. This paper presents a method for solving the problem by the former approach. We propose a systematic method for incorporating the tradeoff wisdom inspired by the circuit domain knowledge in the formulation of the composite cost function. Key issues have been identified and the problem has been divided into two parts: a) normalization of objective functions and b) assignment of weights to objectives in the cost function. A nonlinear, parameterized normalization strategy has been proposed and has been shown to be better than traditional linear normalization functions. Further, the designers' problem specific knowledge is assembled in the form of a partially ordered set, which is used to construct a hierarchical cost graph for the problem. The scalar cost function is calculated based on this graph. Adaptive mechanisms have been introduced to dynamically change the structure of the graph to improve the chances of reaching the near-optimal solution. A correlated double sampling offset-compensated switched capacitor analog integrator circuit and an RF low-noise amplifier in an industry-standard 0.18mum CMOS technology have been chosen for experimental study. Optimization results have been shown for both the traditional and the proposed methods. The results show significant improvement in both the chosen design problems     

Characterizing conflicts in fair division of indivisible goods using a scale of criteria

We investigate five different fairness criteria in a simple model of fair resource allocation of indivisible goods based on additive preferences. We show how these criteria are connected to each other, forming an ordered scale that can be used to characterize how conflicting the agents’ preferences are: for a given instance of a resource allocation problem, the less conflicting the agents’ preferences are, the more demanding criterion this instance is able to satisfy, and the more satisfactory the allocation can be. We analyze the computational properties of the five criteria, give some experimental results about them, and further investigate a slightly richer model with \(k\)-additive preferences.     

Positional Voting Rules Generated by Aggregation Functions and the Role of Duplication

In this paper, we consider a typical voting situation where a group of agents show their preferences over a set of alternatives. Under our approach, such preferences are codified into individual positional values, which can be aggregated in several ways through particular functions, yielding positional voting rules and providing a social result in each case. We show that scoring rules belong to such class of positional voting rules. But if we focus our interest on OWA (ordered weighted averaging) operators as aggregation functions, other well‐known voting systems naturally appear. In particular, we determine those ones verifying duplication (i.e., clone irrelevance) and present a proposal of an overall social result provided by them.     

A multi-criteria sorting procedure with Tchebycheff utility function

In this study, a Tchebycheff utility function based approach is proposed for multiple criteria sorting problems in order to classify alternatives into ordered categories, such as A, B, C, etc. Since the Tchebycheff function has the ability to reach efficient alternatives located even in the non-convex part of the efficient frontier, it is used in the proposed sorting approach to prevent such alternatives being disadvantages. If the preferences of the DM are not exactly known, each alternative selects its own favorable weights for a weighted Tchebycheff distance function. Then, each alternative is compared with the reference alternatives of a class to compute its strength over them. The average strengths are later used to categorize the alternatives. The experimental analysis results on the performance of the algorithm are presented.     

The specification of distributed sensing and control

Logical Sensor System Specification (LSS) has been introduced as a convenient means for specifying multi-sensor systems and their implementations. In this article we demonstrate how control issues can be handled in the context of LSS. In particular, the Logical Sensor Specification is extended to include a control mechanism which permits control information to (1) flow from more centralized processing to more peripheral processes, and (2) be generated locally in the logical sensor by means of a micro-expert system specific to the interface represented by the given logical sensor. Examples are given including a proposed scheme for controlling the Utah/MIT dextrous hand.     

A distributed logic for Networked Cyber-Physical Systems

We describe a distributed logical framework designed to serve as a declarative semantic foundation for Networked Cyber-Physical Systems. The framework provides notions of facts and goals that include interactions with the environment via external goal requests, observations that generate facts, and actions that achieve goals. Reasoning rules are built on a partially ordered knowledge-sharing model for loosely coupled distributed computing. The logic supports reasoning in the context of dynamically changing facts and system goals. It can be used both to program systems and to reason about possible scenarios and emerging properties.     

Stakeholder ranking of watershed goals with the vector analytic hierarchy process: Effects of participant grouping scenarios

More effective methods of eliciting and summarizing stakeholders' goals can assist in improving watershed management. This paper discusses the process of summarizing the goals that were generated during a workshop of watershed stakeholders in Virginia by using the Vector Analytic Hierarchy Process, and then grouping them into homogeneous subgroups by using two different methods: 1) assigning subgroups based on individuals' stated affiliations from a participant bio-sheet; and 2) assigning subgroups based on the similarity of individuals' actual preferences between the goals. Several different clustering approaches are considered for creating the preference-based subgroups, and the relative advantages and disadvantages of each approach are discussed. The process of combining the subgroups to generate a single overall preference structure for the group as a whole is also considered, and the final results are compared based on both the resulting rankings and the coherence, or variability of opinion, that they reflect. Determining the “best” set of subgroups can be valuable not only in exploring the underlying nature of the population's preferences, but also in supporting additional discussion and analysis of the results. As such, it can ultimately lead to much stronger and better informed decision-making by the stakeholders.     

Towards Model-based Methods for Developing Model-based Systems

Model-based reasoning (MBR) is a means of reasoning about models of all kinds, as appropriate to the task at hand. This includes adaptation of models in response to changes in a problem-solving context or task goals. Thus, MBR exemplifies the characteristics of a smart adaptive system. Constructing appropriate models and matching them to the best inference engines requires a means of describing or defining models. This can be achieved by means of a set of generic model properties. As well as defining models one may decompose the problem domain into a number of tasks to be performed. It then becomes possible to design appropriate models for the domain by mapping these tasks to the model properties. This mapping can be instantiated procedurally, however more generality will result if a model-based approach is taken. As a step towards that goal quality function deployment is investigated as a suitable design method.