Multi attribute architecture design decision for core asset derivation

Abstract

Software Product Line (SPL) is an effective approach in software reuse in which core assets can be shared among the members of the product line with an explicit treatment of variability. Core assets, which are developed for reuse in domain engineering, are selected for product specific derivation in application engineering. Decision making support during product derivation is crucial to assist in making multiple decisions during product specific derivation. Multiple decisions are to be resolved at the architectural level as well as the detailed design level, address the need for assisting the decision making process during core asset derivation. Architectural level decision making is based on imprecise, uncertain and subjective nature of stakeholder for making architectural selection based on non- functional requirements (NFR). Furthermore, detail design level involves the selection of suitable features which have the rationale behind each decision. The rationale for the selection, if not documented properly, will also result in loss of tacit knowledge. Therefore, a multi-attribute architecture design decision technique is proposed to overcome the above mentioned problem. The technique combines Fuzzy Analytical Hierarchy Process (FAHP) with lightweight architecture design decision documentation to support the decision making during core asset derivation. We demonstrate our approach using the case study of Autonomous Mobile Robot (AMR). The case study implementation shows showed that the proposed technique supports software engineer in the process of decision making at the architecture and detail design levels.