Publication at Faculty of Mathematics and Physics |
2006
Abstract
In this paper we address the problem of encoding of planspace planning problems as a CSP. We propose a constraint model for expressing plan-space partial plans.
As an enrichment of the notion of partial plan we propose a world state management within that. We incorporate this world state management into our CSP model too.
Next we propose an algorithm for solving this CSP model. The algorithm builds the constraint model dynamically as conflicts are resolved in the model.
The performed experiments showed the usefulness of state management in the constraint model. The usage of state management allows the additional search space pruning compared to the model without states.