Reinforcement Learning Methods on Optimization Problems of Natural Gas Pipeline Networks

Traditional optimization methods of transport and distribution of natural gas pipeline networks have been widely used up to now with some problems in efficiency, cost, and flexibility, which are hard to be solved in the framework of traditional methods. In order to find the optimal solution in the constraints of each target of this optimization problem, this paper establishes a simulation model based on a part of a natural gas pipeline networks, and utilizes the reinforcement learning (RL) algorithm to analyze the model. The challenge of sparse rewards will also be dealt with. Then the optimal strategy of transport and distribution of gas in this model is obtained with different demands and initial conditions. The parameters of the operation of the strategy can be displayed in the simulation model and its advantages can also be fully reflected. Therefore, the scheme proposed in this paper can be directly or indirectly applied to the practical natural gas transport process.