Metaheuristic Search Approach Based on In-house/Out-sourced Strategy to Solve Redundancy Allocation Problem in Component-Based Software Systems

The redundant components' nodes are widely used as design tactic to improve the reliability of complex embedded systems. The challenge is that; the obtained solution might not be sufficient because it demands high cost, and the system might be failed if the predicted cost does not estimated reasonably. Many publications have addressed this area to compose the systems by assembling COTS components only. However, combination of outsourced components and in-house developed components is essential to produce cost effective systems. This paper proposes an architecture optimization approach based on Metaheuristic Swarm Intelligent algorithm in order to answer the question: What type of components to be selected and how many components are appropriate for each subsystem in order to obtain optimal software architecture. The objective is to minimize cost under reliability constraints by combining components from different sources in order to compose system that satisfies the requirements. Sensitivity analysis applied on a numerical simple case study showed the potentiality of the approach.