A new algorithm for pure 0–1 linear programs with inequality constraints

More and more attention has been paid to the Hopfield neural network (HNN) algorithm since it was used by Hopfield and Tank to solve TSPs in 1985. However, the algorithm can only solve the pure 0–1 linear programs with equality constraints, and many tests showed that it is unreliable. This paper proposes an HNN-type algorithm which can solve some pure 0–1 linear programs with inequality constraints and guarantee to find their ‘near-optimal’ solutions. Finally, an example is given.