Resource sharing linear logic
Abstract In this paper, we introduce a new logic that we call ‘resource sharing linear logic (RSLL)’. In linear logic (LL), formulas without modality express some resource-conscious situation (a formula can be used only once); formulas with modality express a situation with unlimited resources. We introduce the logic RSLL in which we have a strengthened modality (S5-modality) that can be understood as expressing not only unlimited resources but also resources shared by different agents. Observing that merely strengthening the modality allows weakening axiom to be derivable in a Hilbert-style formulation of this logic, we reformulate RSLL as a logic similar to affine logic by a hypersequent calculus that has weakening as a primitive rule. We prove the completeness of the hypersequent calculus with respect to phase semantics and the cut-elimination theorem for the system by a syntactical method. We also prove the decidability of RSLL via a computational interpretation of RSLL, which is a parallel version of Kopylov’s computational model for LL. We then introduce an explicit counterpart of RSLL in the style of Artemov’s justication logic (JRSLL). We prove a realization theorem for RSLL via its explicit counterpart.