Risk based incremental survivable network design

This paper presents a new approach for incremental survivable network design based on the use of risk analysis techniques. The objective of the design approach is: given a fixed budget, determine how to incrementally improve the network survivability in order to reduce the risk from network failures. Risk is defined as the product of the failure probability and the damage resulting from the failure. The design approach consists of two parts: a risk assessment and a risk reduction investment strategy. Fault tree models, which depict causal relationships among failure events in the network are used for the risk assessment. The risk-reduction investment strategy is used to determine an allocation of budget for implementing a survivability technique (e.g., link protection) in different parts of the network to minimize the network risk. Mixed Integer Linear Programming (MILP) formulations and greedy-based heuristics are proposed for solving the minimum-risk design problems. Numerical results illustrating the investment strategy for link and path protections, using an MILP approach and proposed heuristic algorithms, along with the comparisons of different capital investment alternatives on the basis of risk consideration are presented and discussed.