This paper describes recent efforts to incorporate integer programming capability within nonlinear programming algorithms. The structure of problems may be exploited in producing an integer feasible solution adjacent to a local continuous optimum. This solution can be accepted as near-optimal if it is sufficiently close in objective value to the continuous optimum, or else be used as an enhanced bound in a branch and bound procedure. Application areas of current interest to which these algorithms have been applied include asset allocation models for portfolio management, and yield management with side conditions. Computational experience in the application of these models is discussed.