Contingencies are unexpected crises or events that cause a major threat to the

safety, security and well-being of a certain population. This research effort builds

upon the work on contingency logistics reliability models by Miman (2008) who

extended the preliminary work conducted by Thomas (2004) that provides the

modeling approach which takes a mission success orientation and focuses on the

ability to recover from or prevent a contingency logistics failure. Miman (2008)

proposes the sustainability model of a contingency logistics network using the

concept of selective maintenance. This problem, once formulated, is a non-convex,

non-linear, non-separable, multi-dimensional, discrete knapsack problem. These

problems are known to be NP hard. Therefore, one needs to explore heuristic

solutions in search of robust and effective solution approaches. He developed a

memetic algorithm, GAFTS, and proposed this for identifying the best set of

maintenance actions to sustain the contingency logistics network. Besides, he used

Physical Programming, a multi criteria optimization procedure, to exploit a network

manager’s preference toward the numerous criteria (reliability, cost, time, resource

utilization etc...) judiciously.

This research effort continues the exploration of heuristic techniques for the

sustainability model developed by Miman (2008) and develops a hybrid heuristics

technique, EDGASA, incooperating simulating annealing (SA) procedure with

genetic algorithm (GA). Comparisons of EDGASA with GA and SA reveal that it

outperforms in terms of average reliability, best reliability and worst reliability found

at an expense of increased solution time.

One of the contributions of this study is a multi-objective modeling approach

developed based on utopia distance that aims at minimizing the weighted distance

between a solution to the ideal point that could be achieved. The study fills some of

the voids in the contingency logistics networks’ solution and modeling and highlights

potential studies by applying the hybrid heuristic developed as well as multiobjective

modeling approach proposed to other problems.