Reoptimize

Reoptimize is the act of performing an optimization anew in response to changes in problem data, constraints, or objectives. It occurs when an initial solution may no longer be feasible or near-optimal due to new information, disturbances, or evolving requirements. Reoptimization often seeks to reuse information from the original solve to speed up finding a new solution.

In practice, reoptimization uses techniques such as warm starts, incremental updates, sensitivity analysis, and parametric optimization. Depending on the problem class, the previous solution can be adjusted minimally, only the affected components re-solved, or the problem reformulated with recourse variables. The goal is to preserve solution quality while reducing computation time compared with solving from scratch.

Applications span operations research, computer science, and engineering. In supply chain management, inventory and routing decisions may be reoptimized as demand or costs change. In scheduling and manufacturing, reoptimization supports dynamic replanning. In energy systems and telecommunications, real-time reoptimization helps adapt to fluctuations and failures.

Related concepts include sensitivity analysis, dynamic optimization, and recourse in stochastic programming. The difficulty of reoptimization depends on problem structure; some changes permit quick updates, while others require a full re-solve. In all cases, reoptimization aims to maintain or improve decision quality with reduced computational effort.