 |
Center for Nonlinear Studies |
A natural disaster and certain kinds of weapons may cause extensive damage to a network, such as a power grid. During recovery operations, it is desirable to re-install the network with the least cost, where the cost of installing a node depends on which other nodes have already been installed (those could act as supply nodes). Finding the optimal installation sequence turns out to be a hard optimization problem over the n! permutations. Fortunately, there exist efficient approximation algorithms as well as exact solutions using dynamic programming. Many open problems remain in this area, but it is clear that optimization of network recovery operations would save the country a lot of resources in future disaster situations. (joint work with Milan Bradonjic, T-5)
Host: Peter Loxley, loxley@lanl.gov