Estimation of Container Ship Principal Dimensions Based on Key Factors QRNN Algorithm

At present, the ship’s principal dimensions are mainly determined by the parent ships or the empirical formula method. The amount of calculation is large and complicated, and there is no integrated method of estimating the principal dimensions. Aiming at the estimation of principal dimensions of container ships, a Quantile Regression Neural Network (QRNN) based on key factors is established, and the sample model is trained through the collected data, thereby reducing the error rate. Through this algorithm, the estimated models for ship length, length between perpendiculars, moulded width, moulded depth and average draught are obtained. The error rate is all less than 6.8%, and the error rate is smaller than traditional algorithm. Applying the idea of QRNN based on key factors algorithm to the estimation of the principal dimensions of container ships can provide designers with certain reference opinions during the preliminary design of the ship.