Discrete Optimizing Solution Procedures for Linear and Nonlinear Integer Programming Problems

We present a method for approximating the solution of mixed integer non-concave programming problems in bounded variables. We present computational results for 39 test problems which suggest that the procedure offers a practical useful way of approximating solutions of programming problems of the type tested. We have also discovered an apparent regularity in the distribution of local optima generated by the method; the method seems to generate Beta distributions in all problems for which results are available. This suggests interesting opportunities for further exploration.