A new heuristic algorithm for probabilistic optimization

A new heuristic method is presented for solving a class of constrained probabilistic optimization problems where the random coefficients involved in the objective function are continuous. This optimization technique is general purpose and computationally feasible. The heuristic idea here has emanated from an analogy of the expected value function with the centre of gravity of a linear physical system. Theoretical results on the efficiency and reliability of the involved algorithm are presented along with computational results on some standard test problems and a real-world problem of antibody inventory control. The importance and scope of the approach is highlighted.