Game Model And Throughput Forecast Based On Neural Network Of Container Port

Because of the drastic competition among the container port world widely, the trend of the container transportation and the structure of the market of container port have been analyzed in detail. Considering the characters of the port industry, which forms the oligopoly market, its demand and supply model is given, and the effects of many competition strategies such as pricing, service, superfluous capacity is analyzed, which is help to develop the port's competition capability, perfect the port planning.Then, based on the game theory, a port game model is established to simulate the competition among the container port. In order to sum up the rule of the port competition, several types have been simulated under different conditions, and their Nash equilibriums have been gained, which is very important to the port in the reality.There are many factors influence the throughput of container port, so the artificial neural network (ANN) is a more suitable method to forecast the throughput than the tradition ones. In this dissertation, an improved BP neural network model and a recurrent neural network (RNN) model is founded to forecast the Dalian port's throughput. According to the historical data and the forecast results, the two models both have good performance, especially the latter.In the end, the economy, hinterland, nature, management, competition environment, and other conditions of Dalian port is discussed. Combined with the result of the forecast, the experience of the big container port, some advice has been bring forward to Dalian port, which could promote the development of district economy and vice versa.