A Separation Logic for Resource Distribution
Résumé
We define a separation logic (BI-Loc) that is an extension of the Bunched Implications (BI) logic with a modality for locations. Moreover, we propose a general data structure, called resource tree, that is a node-labelled tree in which nodes contain resources that belong to a partial monoid. We also define a resource tree model for this logic that allows to reason and prove properties on resource trees. We study the decidability by model checking of the satisfaction and the validity in this separation logic and also introduce a sequent calculus for deciding validity by deduction w.r.t. a resource model. Then, we relate the separation logic and resource trees to some applications and finally define a sequent calculus for BI-Loc dedicated to a theorem proving approach.