Abstract: Under the background of "Internet +",many manufacturers provide online custom business to increase the sales revenue and improve the competitive ability.In order to cope with the requirements of the customers about the high quality,fast delivery time and the randomness of outputs of each stage about the line,the manufacturer can design the management mechanism for the workshop learning from the model of Amoeba Operation.According to the product process,the workshop is divided into multiple organizations of amoeba that are relatively independent.The adjacent amoebas decide the price of the materials and trade with each other orienting to the market and the orders.When the promised delivery time affects the passing rate and total costs,for the decentralized decision-making,the Stackelberg game model based on the inadequate outputs penalty mechanism is set up including decisions of the input quantities and promised delivery time.For the centralized decision-making,a nonlinear optimization model with multiple decision variables is set up based on mechanism of resources sharing,the existence and uniqueness of the optimal solution is proved by establishing the HESSE matrix.The performance distribution and interests incentive can be realized for all the amoeba organizations based on the mechanism of sharing of punishment and procurement costs.Based on the case study and parameters sensitivity,the main conclusions are as follows. With the increase of input quantities, the profits of the amoebas and the overall workshop first increase and then decline;The price discount for the delayed delivery is greater,the previous amoebas of the line would be more favorable but the last amoeba and the overall workshop would be more unfavorable. The greater the volatility of the passing rate is, the lower the expectd profits for each amoeba and the overall workshop is. The theoretical and practice values of the paper are as follows:the design of management mechanism about the workshop according to the Amoeba Operation; the consideration of the factors influencing the passing rate;the optimization and coordination of the n-level supply chain.

Key words: online custom demands, random outputs, amoeba operation, desicion of input quantities, innovation of management mechanism, performance incentives