Reputation: 41817
SAT solvers can be used to solve the traveling salesman problem, where the sum of costs between chosen edges is important. I understand that you then ask the solver to go again finding a lesser cost. How is that maximum represented in conjunctive normal form?
Upvotes: 0
Views: 139
Reputation: 466
The idea is to translate pseudo-Boolean constraints into CNF and many different encodings are available. A naive encoding is to enumerate all partial models that would result in a higher cost (which is exponential).
Note: I am co-author of the following publication: You may find the PBLib useful as it provides different encodings.
Upvotes: 1