Heuristics for NP-hard optimization problems - simpler is better!?

Abstract

We provide several examples showing that local search, the most basic metaheuristics, may be a very competitive choice for solving computationally hard optimization problems. In addition, generation of starting solutions by greedy heuristics should be at least considered as one of very natural possibilities. In this critical survey, selected examples discussed include the traveling salesman, the resource-constrained project scheduling, the channel assignment, and computation of bounds for the Shannon capacity.

Keywords

• > optimization
• metaheuristics
• local search
• greedy construction
• traveling salesman problem