Traditional methods usually formulate and solve the integrated multilevel production planning problems with hierarchical models. Theoretically, the approaches of using a single model to describe and optimize the problem could reach much better performance, although it is quite difficult to realize the optimization process. Based on the JIT idea, a new objective function for optimization is proposed in this paper, and a single model is established which simultaneously formulates and optimizes the lot sizing and scheduling problem. A genetic algorithm is designed for solving this problem and the results are satisfying.