 |
Center for Nonlinear Studies |
As we begin to study more complex logistical problems the weaknesses of various optimization techniques become very restricting. Hybrid optimization methods attempt to solve this weakness by combining the strengths of different optimization methods. This talk will give a lighting fast overview of several optimization techniques including Mixed Integer Programming, Constraint Programming, and Local Search, and introduce Comet, a cutting-edge modeling language for hybrid optimization. We will demonstrate how hybrid optimization was used to solve a complex disaster recovery problem, the Single Commodity Allocation Problem (SCAP). SCAPs investigate how to store a single commodity throughout a populated area to minimize its delivery time after a disaster has occurred. These are complex stochastic optimization problems that combine resource allocation, warehouse routing, and multiple vehicle routing. We present a hybrid-optimization algorithm for solving this problem to aid policy makers in preparing for and recovering from real-world disasters.
Host: Russell Bent