Verification and Uncertainties in Self-integrating System

Abstract

Self-integrating and self-improving system are required to verify their state in order to understand whether they have achieved their goal or need to adapt themselves to reach it. In this short position paper, we outline the main challenges specifically when verifying systems interacting with each other and operating under uncertainties. A short outline of the uncertainties is given as well as a brief roadmap to overcome the main challenges faced by autonomous interwoven systems, operating in an open world with incomplete knowledge.

Towards Reactive Planning with Digital Twins and Model-Driven Optimization

“…Usually, such approaches separate the planning which happens offline and the execution of a plan which happens online while the system is running. This however neglects the online uncertainty of changes that might happen in the system in parallel to the realization of the plan [14]. As a result of these changes, (i) parts of the plan may not be executable any more, or (ii) realizing the initial plan leads to an inferior quality in the goal state of the system.…”