Study On The Optimization Of A Production And Sales System

The upstream and downstream enterprises in the supply chain can be seen as the relationship between the buyer and the seller.It will lead to imbalance between supply and demand which the upstream enterprise does not take into account the needs of the downstream enterprise.Therefore,the study of production and sales between the upstream and downstream enterprises is important for the actual supply chain,which can achieve the long-term operation of each enterprise' the highest profits.First this thesis establishes a production and sale model based on one retailer' demand-driven.The system includes a producer and a retailer,which have independent inventory and decision-making process.The producer in the model can be considered as the demand-driven conveyor-serviced production station(CSPS),meanwhile the inventory status of the bank is influenced by the retailer's replenishment order and its own production capacity.The retailer takes a dynamic replenishment strategy,which the inventory status is influenced by the random customer's needs and the actual replenishment of the producer.The control variable of the producer is the look-ahead distance of the processing site and the optimization goal is to find the optimal control strategy of look-ahead distance.Considering the random part arrival,random processing time in the system,the thesis adopts a model-independent learning algorithm to solve the optimal or sub-optimal control strategy of the producer.The retailer's control variable is the replenishment point,and the optimization goal is to find the optimal inventory control strategy.Due to the random customer demand,the thesis uses the reinforcement learning algorithm to solve the retailer's optimal control problem.Considering the fact that the producer usually supply a number of retailers,the thesis further studies the optimization control of a production and sales system based on multiple retailers' demand-driven.There is a static game relationship between multiple retailers in the system,that is,the replenishment of each retailer is time sequence,and the retailers do not know the replenishment strategy of the rest of the retailers.After the producer replenishes to the retailer,the status of the producer' bank has changed.Finally the thesis uses the reinforcement learning algorithm to solve the producer and the retailers optimal control problem and analyzes the game relationship between multiple retailers from customer arrival rate,warehouse capacity and replenishment period.