Abstract. The goal of this roadmap paper is to summarize the state-ofthe-art and to identify critical challenges for the systematic software engineering of self-adaptive systems. The paper is partitioned into four parts, one for each of the identified essential views of self-adaptation: modelling dimensions, requirements, engineering, and assurances. For each view, we present the state-of-the-art and the challenges that our community must address. This roadmap paper is a result of the Dagstuhl Seminar 08031 on "Software Engineering for Self-Adaptive Systems, " which took place in January 2008.

No abstract

This paper reviews the current state of the art of requirements engineering (RE) research and identifies RE research challenges for future systems. First, the paper overviews the highlights of RE research over the past two decades; the research is considered with respect to requirements technologie, including notations and methodologies, developed to address specific RE tasks, such as elicitation, modeling, and analysis. Such a review enables us to identify mature areas of research, as well as areas that warrant further investigation. Next, we identify several research challenges posed by emerging systems for the future. In order to help delineate the scope of future RE research directions, we then identify several strategies for performing RE research. (The spectrum of research strategies ranges from empirical research to paradigm shifts.) Finally, within the context of these RE research strategies, we identify "hot areas" of research that address RE needs for emerging systems of the future.

No abstract

Self-adaptive systems have the capability to autonomously modify their behaviour at run-time in response to changes in their environment. Self-adaptation is particularly necessary for applications that must run continuously, even under adverse conditions and changing requirements; sample domains include automotive systems, telecommunications, and environmental monitoring systems. While a few techniques have been developed to support the monitoring and analysis of requirements for adaptive systems, limited attention has been paid to the actual creation and specification of requirements of self-adaptive systems. As a result, self-adaptivity is often constructed in an ad-hoc manner. In this paper, we argue that a more rigorous treatment of requirements explicitly relating to self-adaptivity is needed and that, in particular, requirements languages for self-adaptive systems should include explicit constructs for specifying and dealing with the uncertainty inherent in self-adaptive systems. We present RELAX, a new requirements language for selfadaptive systems and illustrate it using examples from the smart home domain.

Embedded systems are pervasive and frequently used for critical systems with time-dependent functionality. Dwyer et al. have developed qualitative specification patterns to facilitate the specification of critical properties, such as those that must be satisfied by embedded systems. Thus far, no analogous repository has been compiled for realtime specification patterns. This paper makes two main contributions: First, based on an analysis of timing-based requirements of several industrial embedded system applications, we created real-time specification patterns in terms of three commonly used real-time temporal logics. Second, as a means to further facilitate the understanding of the meaning of a specification, we offer a structured English grammar that includes support for real-time properties. We illustrate the use of the real-time specification patterns in the context of property specifications of a real-world automotive embedded system.

Self-adaptive systems have the capability to autonomously modify their behavior at run-time in response to changes in their environment. Self-adaptation is particularly necessary for applications that must run continuously, even under adverse conditions and changing requirements; sample domains include automotive systems, telecommunications, and environmental monitoring systems. While a few techniques have been developed to support the monitoring and analysis of requirements for adaptive systems, limited attention has been paid to the actual creation and specification of requirements of self-adaptive systems. As a result, self-adaptivity is often constructed in an ad-hoc manner. In order to support the rigorous specification of adaptive systems requirements, this paper introduces RELAX, a new requirements language for selfadaptive systems that explicitly addresses uncertainty inherent in adaptive systems. We present the formal semantics for RELAX in terms of fuzzy logic, thus enabling a rigorous treatment of requirements that include uncertainty. RELAX enables developers to identify uncertainty in the requirements, thereby facilitating the design of systems that are, by definition, more flexible and amenable to adaptation in a systematic fashion. We illustrate the use of RELAX on smart home applications, including an adaptive assisted living system. IntroductionAs applications continue to grow in size, complexity, and heterogeneity, it becomes increasingly necessary for computing-based systems to dynamically self-adapt to changing environmental conditions. We call these systems dynamically adaptive systems (DASs). Example applications that require DAS capabilities include automotive systems, telecommunication systems, environmental monitoring, and power grid management systems. The distributed nature of DASs and changing environmental factors (including human interaction) make it difficult to anticipate all the explicit states in which the system will be during its lifetime. As such, a DAS needs to be able to tolerate a range of environmental conditions and contexts, but the exact nature of these contexts remains imperfectly understood. One overarching challenge in developing DASs, Address(es) of author(s) should be given

An algorithm for the rapid analytical determination of the accessible surface areas of solute molecules is described. The accessible surface areas as well as the derivatives with respect to the Cartesian coordinates of the atoms are computed by a program called "MSEED," which is based in part on Connolly's analytical formulas for determining surface area. Comparisons of the CPU time required for MSEED, Connolly's numerical algorithm DOT, and a program for surface area determination (ANA) based on Connolly's analytical algorithm, are presented. MSEED is shown to be as much as 70 times faster than ANA and up to 11 times faster than DOT for several proteins. The greater speed of MSEED is achieved partially because nonproductive computation of the surface areas of internal atoms is avoided. A sample minimization of an energy function, which included a term for hydration, was carried out on MET-enkephalin using MSEED to compute the solvent-accessible surface area and its derivatives. The potential employed was ECEPPiZ plus an empirical potential for solvation based on the solvent-accessible surface area of the peptide. The CPU time required for 150 steps of minimization with the potential that included solvation was approximately twice as great as the CPU time required for 150 steps of minimization with the ECEPP/2 potential only.