Abstract: Taking the 3-level multi-product supply chain as an example，the supply chain was ran by using four strategies， i.e.， Push， Pull， hybrid Push & Pull and improved Push. The global uncertainty of supply chain was quantified by total entropy ratio. Genetic algorithm(GA) was combined with simulation-based optimization(SBO) method to deal with the difficulty of large amounts of computation and uncertainty. The gain in the control law being as the decision variable， the uncertainty was optimized， and other common performance indicators such as customer satisfaction， excess inventory， delayed delivery and total cost were calculated under the optimal decision variables. The simulation results showed that when the demand exceeds supply， the hybrid Push & Pull strategy can reduce the total cost. When the supply exceeds demand， the improved Push strategy can minimize the uncertainty of the supply chain.

Key words: supply chain, uncertainty, information entropy, Pull/Push strategy, SBO