The Complexity of Finite Memory Programs with Recursion

Authors

  • Neil D. Jones
  • Steven S. Muchnick

DOI:

https://doi.org/10.7146/dpb.v5i68.6486

Abstract

In an earlier paper (JACM, 1976) we studied the computational complexity of a number of questions of both programming and theoretical interest (e.g. halting, looping, equivalence) concerning the behaviour of programs written in an extremely simple programming language. These finite memory programs or fmps model the behaviour of FORTRAN-like programs with a finite memory whose size can be determined by examination of the program itself.

The present paper is a continuation in which we extend the analysis to include ALGOL-like programs (called fmp^(rec) s) with the finite memory augmented by an implicit pushdown stack used to support recursion.

Our major results are the following. First, we show that at least deterministic exponential time is required to determine whether a program in the basic fmpr~C model accepts a nonempty set. Then we show that a model with a limited version of call-by-name requires exponential space to determine acceptance of a nonempty set, and that a more sophisticated model with rewritable conditional formal parametershas an undecidable halting problem. The same lower bounds apply to the equivalence problem, which in contrast to the situation for the basic fmp model is not known to be decidable (since it is not known whether equivalence of deterministic pushdown automata is decidable).

Author Biographies

Neil D. Jones

Steven S. Muchnick

Downloads

Published

1976-12-01

How to Cite

Jones, N. D., & Muchnick, S. S. (1976). The Complexity of Finite Memory Programs with Recursion. DAIMI Report Series, 5(68). https://doi.org/10.7146/dpb.v5i68.6486