Linear Tree Constraints
share
Linear tree constraints were introduced by Hofmann and Rodriguez in the context of amortized resource analysis for object oriented programs. More precisely, they gave a reduction from inference of resource types to constraint solving. Thus, once we have found an algorithm to solve the constraints generated from a program, we can read off the resource consumption from their solutions. These constraints have the form of pointwise linear inequalities between infinite trees labeled with nonnegative rational numbers. We are interested in the question if a system of such constraints is simultaneously satisfiable. Bauer and Hofmann have recently identified a fragment of the tree constraint problem (UTC) that is still sufficient for program analysis and they proved that the list case of UTC is decidable, whereas the case with trees of degree at least two remained open. In this paper, we solve this problem. We give a decision procedure that covers the entire range of constraints needed for resource analysis.
READ FULL TEXT