A history based logic for dynamic preference updates (Q2209166)

In the introductory section, the authors motivate their approach as follows: ``In this work, we extend history based structures by introducing preference modalities -- first a static one followed by a dynamic one. By achieving this, we relate the subject to the domains of preference logic and distributed multi-agent systems. After noting the fundamental properties of the preference modality, we proceed to describe various strategies to dynamically update preferences. Following, we focus on some immediate applications of our models and conclude by pointing out the potential use of this framework in various game theoretical situations.'' Under this approach, history based models are considered to be the basic logical structure. This structure employs a language with a knowledge operator, a temporal next-time operator, and the until operator. The authors aim at introducing subjective preferences to the history based models. To this effect, the syntax of the logic of history based structures is augmented with a modal operator \(\Diamond_i\). The formula \(\Diamond_i \varphi\) means that there is a history which is at least as good as the current one and satisfies \(\varphi\) for agent \(i\). There is a preference relation \(\preceq_i\) in the semantic models, which is a pre-order satisfying reflexivity and transitivity (not necessarily total). For an agent \(i\), and possibly infinite histories \(H\), \(H^{\prime}\), the expression \(H \preceq_i H^\prime\) means: ``the agent \(i\) (weakly) prefers \(H^\prime\) to \(H\)''. The preference modality obtain then the following definition: \(H , t \vDash \Diamond_i \varphi\) iff \(\exists H^\prime. H \preceq_i H^\prime\) and \(H^\prime, t \vDash \varphi\). In this way a history based preference logic \(\mathsf{HBPL}\) is constructed.