Multi-Criteria Decision-Making (MCDM) Approach for Software Architecture Selection in Cloud Computing Using Evidential Reasoning and Bayesian Inference Techniques

Choosing the optimal software architecture for cloud-based systems is a critical and complex Multi-Criteria Decision Making (MCDM) problem, characterized by multiple, often conflicting, and interdependent criteria such as performance, cost, scalability, deployment speed, security, and maintainability.This research addresses this challenge by proposing and applying an integrated MCDM methodology that leverages Evidential Reasoning (ER) and Bayesian Inference (BI).The study's primary objective is to provide a robust and transparent framework for evaluating six common architecture styles: Monolithic, Microservices, Layered, Serverless, Event-Driven, and Service-Oriented Architecture (SOA).The methods employed involved a multi-stage process.First, criteria weights were derived using the Analytic Hierarchy Process (AHP) through expert pairwise comparisons.The techniques for handling uncertainty and dependencies were central.ER was utilized to aggregate subjective and objective assessments, representing them as belief distributions to explicitly account for imprecision and ignorance.Concurrently, BI was applied to model probabilistic interdependencies between criteria (Security influencing Performance, Performance influencing Scalability and Cost) within a Bayesian Network.The Intelligent Decision System (IDS) tool facilitated the operationalization of both ER aggregation and Bayesian inference.The results of the AHP weighting revealed the priorities: Performance Republic of South Africa.He is the former National President of the Nigeria Computer Society.