CIAOERP: A Method for Change Impact Analysis and Search-Based Decision Optimization in Cloud ERP Systems Using Feature Forests

Compared with on-premise ERPs, change management is a more serious challenge in cloud ERPs, since changes in an element affect not only other elements but also multiple instances and tenants. Therefore, to make proper decisions, we need a change impact analysis (CIA) method to analyze the way changes are propagated, the affected elements, and the costs imposed. Introducing the concept of feature forest, from multi-software product line engineering, this article addresses how to analyze the impact of changes of “configuration” type in single- and multi-tenant cloud ERPs. Moreover, we propose a multi-objective non-dominated deterministic search algorithm to decide on realizing the changes aiming at minimal change costs. We validated the method by carrying out an empirical experiment following the GQM approach on three real-world cloud ERPs. The evaluation results indicate a decrease of costs of about ۳۳% by the method against the average costs of all possible decisions. Moreover, compared with participants, while showing about ۶۷% similarity, we observed an improvement of about ۴۸% in the costs of decisions.

کلیدواژه ها:

Cloud ERP ، Change Impact Analysis ، Multi Software Product Line ، Feature Forest ، Search-Based Optimization