Multi-Criteria Decision Theory for Enterprise Architecture Risk Assessment: Theory, Modeling and Results

摘要

The paper presents a multiple-criteria decision model based on the concept of pairwise outranking and its use in enterprise architecture assessment. More specifically, the paper applies the Marquis de Condorcet principle found in the ELECTRE models to demonstrate that under highly complex decision situations when existing optimization techniques might not yield optimal outcome, a systematic outranking approach can lead to a satisfactory and safe decision. This paper presents a mathematical version of ELECTRE II (sorting by elimination) implemented in Matlab with visual interpretation. The model was successfully tested on an ex-post notional use case involving four enterprise architecture alternatives based on five evaluation criteria: market uncertainty, technological uncertainty, technical and performance risk, cost and schedule risks, and payoffs and costs.