This paper shows that neighborhood transformations and data-independent order transformations preserve the length of improving paths and order of local optima of neighborhood functions. These results imply that finding effective neighborhood functions for Zero-One IP is at least as hard as finding effective neighborhood functions for any other NPO problem.
- Computational complexity
- Discrete optimization
- Neighborhood functions
ASJC Scopus subject areas
- Management Science and Operations Research
- Industrial and Manufacturing Engineering
- Applied Mathematics