Content area
Full Text
Empir Software Eng (2014) 19:14661501 DOI 10.1007/s10664-014-9305-0
AutoRELAX: automatically RELAXing a goal model to address uncertainty
Erik M. Fredericks Byron DeVries Betty H. C. Cheng
Published online: 12 March 2014 Springer Science+Business Media New York 2014
Abstract Dynamically adaptive systems (DAS) must cope with system and environmental conditions that may not have been fully understood or anticipated during development. RELAX is a fuzzy logic-based specification language for identifying and assessing sources of environmental uncertainty, thereby making DAS requirements more tolerant of unanticipated conditions. This paper presents AutoRELAX, an approach that automatically generates RELAXed goal models to address environmental uncertainty. Specifically, AutoRELAX identifies goals to RELAX, which RELAX operators to apply, and the shape of the fuzzy logic function that establishes the goal satisfaction criteria. AutoRELAX generates different solutions by making tradeoffs between minimizing the number of RELAXed goals and maximizing delivered functionality by reducing the number of adaptations triggered by minor and adverse environmental conditions. In a recent extension, AutoRELAX uses a stepwise adaptation of weights to balance these two competing concerns and thereby further improve the utility of AutoRELAX. We apply it to two industry-based applications involving network management and a robotic controller, respectively.
Keywords Uncertainty Requirements engineering Goal models Genetic algorithms
1 Introduction
A dynamically adaptive system (DAS) may encounter operational contexts that may not have been fully understood or anticipated during requirements elicitation and system
Communicated by: Gordon Fraser and Jerffeson Teixeira de Souza
E. Fredericks ([envelopeback]) B. DeVries B. H. C. Cheng
Department of Computer Science and Engineering, 3115 Engineering Building, 428 S. Shaw Lane, East Lansing, MI 48824-1226, USAe-mail: mailto:[email protected]
Web End [email protected]
B. DeVriese-mail: mailto:[email protected]
Web End [email protected]
B. H. C. Chenge-mail: mailto:[email protected]
Web End [email protected]
Empir Software Eng (2014) 19:14661501 1467
development. Contextual uncertainty (Cheng et al. 2009; Whittle et al. 2009; Baresi et al. 2010; Esfahani et al. 2011) may arise from a combination of unpredictable conditions that can limit a DASs ability to adapt. Uncertainty may be mitigated through self-reconfiguration of a DAS, however introducing adaptation can incur costs in preparing and executing the adaptation. RELAX (Cheng et al. 2009; Whittle et al. 2009) is a requirements specification language for specifying and assessing sources of uncertainty in a DAS. RELAX can be used to make requirements...