Minimum Cost Redundancy Optimization for a Parallel Application on Engineering Systems

ABSTRACT The redundancy technique is widely used in practice to improve performance and enhance the survivability of engineering systems. This paper addresses the problem of minimizing redundancy to satisfy task requirements for a parallel application within engineering systems. According to different task requirements, three redundancy deployment cases with minimum cost are studied. First, we establish an optimization model for redundancy without considering performance demand, and based on this, the case with performance demand is further modeled. Second, we investigate the best allocation policies for a batch of heterogeneous spares while considering performance demand. It is shown that deploying both genuine components and false elements results in a more reliable system. Finally, an example of a control center in the power grid is introduced to illustrate the models, and the results offer useful insights for deciding on redundancy design.