From verification to synthesis under cost-utility constraints

Abstract

Various algorithms and tools for the formal verification of systems with respect to their quantitative behavior have been developed in the past decades. Many of these techniques inherently support the automated synthesis of strategies that guarantee the satisfaction of performance or reliability constraints by resolving controllable nondeterministic choices in an adequate way. More recently, such techniques have been further developed towards the analysis or strategy synthesis under multiple cost and utility constraints. This article deals with Markov decision processes (MDPs) for modeling systems and their environment and provides an overview of recent directions for different types of synthesis problems in MDPs that aim to achieve an acceptable cost-utility tradeoff.

Multi-cost Bounded Tradeoff Analysis in MDP

“…epoch model M 2,1 can only be reached from epochs M i, j with i ≤ 2, j ≤ 1. In our illustration, we only show the transitions between the epoch models that are forwardreachable from M 4,3 ; observe that in this example, these are significantly fewer than what the backward-reachability argument based on the maximum costs gives, which are again only a fraction of all possible epochs.…”

“…The analysis of single-objective (cost-bounded) reachability in MDPs is an active area of research in both the AI and the formal methods communities, and referred to in e.g. [3,18,38,59]. Various model checking approaches for single objectives exist.…”

Breaking the Limits of Redundancy Systems Analysis

“…This enabled the analysis of large-scale systems and speedup their analysis, e.g., for the all-in-one family-based analysis of feature-oriented systems [8,10]. Family-based synthesis using symbolic PMC towards optimal system configurations has been detailed in [3].…”

“…To tackle the aforementioned challenges when modeling and analyzing redundancy systems, we propose to use family-based approaches (see, e.g., [3,8,10,22]) where a single family model comprises the behaviors of all protection combinations. First, such approaches avoid modeling each protection combination individually and to use an automated generation of any combination out from the family model.…”