Resource-Distribution via Boolean Constraint (Extended Abstract)

Proof-search (the basis of logic programming) with multiplicative inference rules, such as linear logic's R and OL, is problematic because of the required non-deterministic splitting of resources. Similarly , searching with additive rules such as NL and R requires a non-deterministic choice between two formulae. Many strategies which resolve such non-determinism, either locally or globally, are available. We present a characterization of a range of strategies for distributing and selecting resources in linear sequent calculus proof-search via a sequent calculus annotated with Boolean constraints. Strategies are characterized by calculations of solutions of sets of Boolean equations generated by searches. Our characterization encompasses lazy (or local), eager (or global) and intermediate (mixed local and global) strategies .