Objectives   In order to improve the positioning ability of the semi-submersible platforms under different sea conditions, three optimal thrust allocation modes based on genetic algorithm were proposed for their dynamic positioning system.

  Methods   By analyzing the influence of different items in the objective function on system power, propeller wear and thrust error, the weight coefficients of the thrust angle and relaxation matrix were determined. Then a limited selection method was proposed to reduce the thrust loss, and the hydrodynamic interference between the propellers was avoided by the method of setting the upstream propellers to be prohibited from entering the thrust forbidden zone. Finally, the three proposed strategies were used for the simulation under complex and variable sea conditions.

  Results   The results show that the limited selection method can effectively reduce the energy consumption, propeller wear and thrust error; the proposed thrust allocation strategy is effective under complex sea conditions.

  Conclusions   The proposed method and strategies can provide ideas and references for the design of dynamic positioning system to a certain extent.