Overview

Self-organizing, adaptive systems (SOAS) are characterized by the fact that they shift decisions regarding the behavior and organizational structure from the designtime to the runtime. That way they gain the possibility to recover from from bad, inappropriate or undesirable conditions caused by a changing environment by the use of adaptation and self-organization mechanisms. This is achieved by an independent reconfiguration of agents or groups of agents resulting in a system structure that adapts to the new requirements of the environmental situation. A change in the environment can be caused by the addition of new tasks or agents up to the loss of capabilities.


The TeSOS project is dedicated to testing SOAS, i.e. the systematic detection of errors. Classically, an error occurs when the actual behavior deviates from the target behavior; the goal of testing is to show/find such failure effects. The adaptive and self-organizing systems underlying mechanisms are now essential to include in the test procedure, as they are able to detect a deviation between good, appropriate or desired behavior and actual behavior. They are then able to reconfigure the system in such a way that it is in a good, suitable or desired state again. For this purpose TeSOS is based on the architecture principle "Corridor Enforcing Infrastructure (CEI)". This spans the space of desired states explicitly as  "Corridor of Correct Behaviour (CCB)", which is monitored at runtime and triggers mechanisms to restore a desired state in the event of deviations. TeSOS aims to prove that the CEI can standardize the testing of SOAS by making the CEI a basic test object and making it testable: The functionality of the individual agents in a SOAS, which is mostly characterized by simple rules, can be tested by classical methods. For adaptivity, on the other hand, the CEI clearly defines what is faulty reconfiguration behavior and which is the most common part of the SOAS that concern adaptation and self-organization, and thus offers the possibility to rely on reconfigurations and to test for adaptation errors. The limitations of the system states for describing the corridor correct behavior form the basis for deriving an automatic test oracle. An approach for isolating the reconfiguration procedures also leads to a generic model-based test architecture that incorporates test case generation, test case execution and test case evaluation and can be used to combine both the test of the whole SOAS as well as system parts, i.e. agents or agent groups, on the basis of CEI.  TeSOS will demonstrate that the CEI is suitable for testing of a SOAS, it is nevertheless not necessary that the CEI was already installed at the time of development; rather it can serve as a guiding principle for testing using the test architecture. During the project we will evaluate the results by the use of five SOAS case studies including a self-organizing production system and an adaptive energy system.

 

Funded by

 

Key Facts

Start date:
01.01.2016
End date:
31.01.2020
Duration:
3 years

 

The pivotal question is how to handle the state space of a self-organizing, adaptive system which is developing in an evolutionary fashion during the execution of the system.

Dr. Hella Ponsar, local head of project

Team

Senior Researcher
Institute for Software & Systems Engineering
Prof. Dr. Alexander Knapp
Professor
Lehrstuhl für Grundlagen Software & Systems Engineering
Director
Institute for Software & Systems Engineering

Institute for Software & Systems Engineering

The Institute for Software & Systems Engineering (ISSE), directed by Prof. Dr. Wolfgang Reif, is a scientific institution within the Faculty of Applied Computer Science of the University of Augsburg. In research, the institute supports both fundamental and application-oriented research in all areas of software and systems engineering. In teaching, the institute facilitates the further development of the faculty's and university's relevant course offerings.

Search