Based on the analysis of inventory multi-objective planning problems under random demand, a multi-objective
(Q, r) model is constructed under three criteria of minimization of the annual expected total relevant cost, the annual expected frequency of stock-out occasions, and the annual expected number of stock-outs quality. To find Pareto front, a hybrid algorithm is designed by using generation algorithm(GA) and differential evolution algorithm to generate non-dominated solutions. Then, an entropy theory-based technique for order preference by similarity to ideal solution is adopted to sort the non-dominated solutions, which gives useful advices for decision makers.