| Summary: | Planning problems are a very important application area for constraint satisfactory problems. Temporal networks are a common way of planning problems representation. In temporal network nodes represent time points and arcs represent time relations. A solution of temporal network means finding such parameters of nodes, that all constraints hold. A nested temporal network, as presented in this thesis, is a special case of the temporal network, which has a restricted structure to be still powerfull for practical use and simultaneously allows using of more effective solving methods. In this thesis nested temporal networks are defined and explored. A polynomial time method for global consistency of alternatives is presented and a proof of NP-completeness for global consistency is demonstrated. If more than one solution exist, it is suitable to have a method for comparing solutions. A common way to accomplish this is price. Usual task is to find solution with minimal price. In the thesis we use the Branch and Bound method with a new filtering algorithm, which exploites the special structure of nested temporal network.
|
|---|
