README

These are C-routines for an implementation of
Lemke's algorithm to solve the LCP  
w = q + Mz >= 0, z >=0,  z^T w=0.

See file sample-input for input.

COMPILE by simply typing

    make

to create the executable file   inlemke

Test with 

    ./inlemke  <  sample-input

d is the covering vector needed to initialize

with   q + dz_0 >= 0,

so it must have a positive entry whenever q has
a negative one.

Restrictions: input are fractions with
int  limitations on size (31 bit).
Big integers should be preferred here.
The gmp library (GNU multiprecision) is much preferable 
for larger LCPs.  The  glemke.c  version has this,
but  gmp  must be installed, and the code adapted.

Bernhard von Stengel 
26 March 2012