We present a new method for the generation of linear invariants which reduces the problem to a non-linear constraint solving problem. Our method, based on Farkas' Lemma, synthesizes linear invariants by extracting non-linear constraints on the coefficients of a target invariant from a program. These constraints guarantee that the linear invariant is inductive. We then apply existing techniques, including specialized quantifier elimination methods over the reals, to solve these non-linear constraints. Our method has the advantage of being complete for inductive invariants. To our knowledge, this is the first sound and complete technique for generating inductive invariants of this form. We illustrate the practicality of our method on several examples, including cases in which traditional methods based on abstract interpretation with widening fail to generate sufficiently strong invariants.

No abstract

We present two algorithms to prove termination of programs by synthesizing linear ranking functions. The first uses an invariant generator based on iterative forward propagation with widening and extracts ranking functions from the generated invariants by manipulating polyhedral cones. It is capable of finding subtle ranking functions which are linear combinations of many program variables, but is limited to programs with few variables. The second, more heuristic, algorithm targets the class of structured programs with single-variable ranking functions. Its invariant generator uses a heuristic extrapolation operator to avoid iterative forward propagation over program loops. For the programs we have considered, this approach converges faster and the invariants it discovers are sufficiently strong to imply the existence of ranking functions.