Loop invariant

From Wikipedia, the free encyclopedia
Jump to: navigation, search

In computer science, a loop is a programming language statement which allows code to be repeatedly executed; an invariant of a loop is a property that holds before (and after) each repetition. It is a logical assertion, sometimes programmed as an assertion. Knowing its invariant(s) is essential for understanding the effect of a loop.

In formal program verification, in particular in the Floyd-Hoare approach, loop invariants are expressed in formal predicate logic and used to prove properties of loops and, by extension, algorithms employing loops (usually correctness properties). A loop invariant should be true on entry into a loop and is guaranteed to remain true after every iteration of the loop. This means that on exit from the loop both the loop invariant and the loop termination condition can be guaranteed.

Because of the fundamental similarity of loops and recursive programs, proving partial correctness of loops with invariants is very similar to proving correctness of recursive programs via induction. In fact, the loop invariant is often the inductive property- the induction hypothesis- one has to prove of a recursive program that is equivalent to a given loop.

Informal example[edit]

The following C subroutine max() returns the maximal value of its argument array a[], provided its length n is at least 1. In line 3, 6, 9, 11, and 13, a property that obviously holds at the respective location has been inserted. The properties in line 6 and 11 agree literally; they are hence an invariant of the loop in lines 5 to 12. When line 13 is reached, that invariant still holds, and it is known that the loop condition i!=n from line 5 must have been false; both properties together imply that m equals the maximum value in a[0...n-1] to be computed by the subroutine, that is, the correct value is returned in line 14.

  1. int max(int n,const int a[n]) {
    
  2.     int m = a[0];
    
  3.     // m equals the maximum value in a[0...0]
    
  4.     int i = 1;
    
  5.     while (i != n) {
    
  6.         // m equals the maximum value in a[0...i-1]
    
  7.         if (m < a[i])
    
  8.             m = a[i];
    
  9.         // m equals the maximum value in a[0...i]
    
  10.         ++i;
    
  11.         // m equals the maximum value in a[0...i-1]
    
  12.     }
    
  13.     // m equals the maximum value in a[0...i-1], and i==n
    
  14.     return m;
    
  15. }
    

Following a defensive programming paradigm, the loop condition i!=n in line 5 should better be modified to i<n, in order to avoid endless looping for illegitimate negative values of n. While this change in code intuitively shouldn't make a difference, the reasoning leading to its correctness becomes somewhat more complicated, since then only i>=n is known in line 13. In order to obtain that also i<=n holds, that condition has to be included into the loop invariant. It is easy to see that i<=n, too, is an invariant of the loop, since i<n in line 6 can be obtained from the (modified) loop condition in line 5, and hence i<=n holds in line 11 after i has been incremented in line 10. However, when loop invariants have to be manually provided for formal program verification, such intuitively too obvious properties like i<=n are often overlooked.

Floyd–Hoare logic[edit]

Specifically in Floyd–Hoare logic,[1][2] the partial correctness of a while loop is governed by the following rule of inference:

\frac{\{C\land I\}\;\mathrm{body}\;\{I\}} {\{I\}\;\mathbf{while}\ (C)\ \mathrm{body}\;\{\lnot C\land I\}}

This means:

  • A while loop does not have the side effect of falsifying I—if the loop's body does not change an invariant I from true to false given the condition C, then I will still be true after the loop has run as long as it was true before.
  • while(C) ... runs as long as the condition C is true—after the loop has run, if it terminates, C is false.

The rule above is a deductive step that has as its premise the Hoare triple \{C\land I\}\;\mathrm{body}\;\{I\}. This triple is actually a relation on machine states. It holds whenever starting from a state in which the boolean expression C\land I is true and successfully executing some program called body, the machine ends up in a state in which I is true. If this relation can be proven, the rule then allows us to conclude that successful execution of the program while (C) body will lead from a state in which I is true to a state in which \lnot C\land I holds. The boolean formula I in this rule is known as the loop invariant.

The following example illustrates how this rule works. Consider the program

while (x < 10)
    x := x+1;

One can then prove the following Hoare triple:

\{x\leq10\}\; \mathbf{while}\ (x<10)\ x := x+1\;\{x=10\}

The condition C of the while loop is x<10. A useful loop invariant I is x\leq10. Under these assumptions it is possible to prove the following Hoare triple:

\{x<10 \land x\leq10\}\; x := x+1 \;\{x\leq10\}

While this triple can be derived formally from the rules of Floyd-Hoare logic governing assignment, it is also intuitively justified: Computation starts in a state where x<10 \land x\leq10 is true, which means simply that x<10 is true. The computation adds 1 to x, which means that x\leq10 is still true (for integer x).

Under this premise, the rule for while loops permits the following conclusion:

\{x\leq10\}\; \mathbf{while}\ (x<10)\ x := x+1 \;\{\lnot(x<10) \land x\leq10\}

However, the post-condition \lnot(x<10)\land x\leq10 (x is less than or equal to 10, but it is not less than 10) is logically equivalent to x=10, which is what we wanted to show.

The loop invariant plays an important role in the intuitive argument for soundness of the Floyd-Hoare rule for while loops. The loop invariant has to be true before each iteration of the loop body, and also after each iteration of the loop body. Since a while loop is precisely the repeated iteration of the loop body, it follows that if the invariant is true before entering the loop, it must also be true after exiting the loop.

Programming language support[edit]

Eiffel[edit]

The Eiffel programming language provides native support for loop invariants.[3] A loop invariant is expressed with the same syntax used for a class invariant. In the sample below, the loop invariant expression x <= 10 must be true following the loop initialization, and after each execution of the loop body; this is checked at runtime.

    from
        x := 0
    invariant
        x <= 10
    until
        x >= 10
    loop
        x := x + 1
    end

See also[edit]

References[edit]

  1. ^ R. W. Floyd. "Assigning meanings to programs." Proceedings of the American Mathematical Society Symposia on Applied Mathematics. Vol. 19, pp. 19–31. 1967. ([1])
  2. ^ Hoare, C. A. R. (October 1969). "An axiomatic basis for computer programming". Communications of the ACM 12 (10): 576–580. doi:10.1145/363235.363259.  edit
  3. ^ Meyer, Bertrand, Eiffel: The Language, Prentice Hall, 1991, pp. 129–131.

Further reading[edit]