MOVES Seminar 21. December 2006

Inductive Decidability Using Implicit Induction Methods

 

 Stephan Falke (University of New Mexico, Albuquerque, NM, USA)

 

 

Abstract:

 

Decision procedures are widely used in automated reasoning tools in order to reason about properties of data structures.  In applications, however, many conjectures fall outside the theory handled by a decision procedure since reasoning about user-defined functions on those data structures is needed.  For this, inductive reasoning has to be employed.

 

Since inductive reasoning typically requires substantial user interaction and guidance it cannot be used as a black-box within verification tools.  In this work we identify classes of function definitions and conjectures about them for which inductive validity can be decided automatically.  For this, we make use of implicit induction methods and decision procedures for the data structures the user-defined functions operate on.

 

The class of equational conjectures considered in this work significantly extends earlier results of Kapur, Giesl and Subramaniam by allowing nonlinear conjectures and by broadening the class of permitted function definitions.