The Northern Sea Route (NSR) is becoming a practical passage for transport during the warm season due to Arctic sea ice reduction. There is a need for new navigational support as LNG shipments are increasing in recent years. In this research, a non-heuristic optimal route search is attempted near the New Siberian Islands. The assumption is that an optimal route is related to the surrounding conditions. Accordingly, an empirical formula considering vessel specifications, and environmental condition parameters can predict transit time along a specific course. Then, a route corresponding to the shortest predicted transit time becomes the optimal route. Multi linear and LASSO regressions are applied over the surroundings of the New Siberian Islands, where three routes lie. While the obtained functions are both overfitting models and not suitable for practical prediction, the LASSO one can estimate a more reasonable range of the standardized velocities. The ranges of the standardized velocities to transit through the domain is from 310 to 1640 km and from 310 to 602 km per day predicted by multi linear and LASSO models, respectively. The LASSO regression analysis might be expected to identify points where all the cautions are necessary along the NSR.

由于北极海冰减少，北海航线（NSR）在温暖季节正成为运输的实用通道。随着近年来液化天然气运输的增加，需要新的导航支持。在这项研究中，尝试在新西伯利亚群岛附近进行非启发式最佳路线搜索。假定最佳路线与周围环境有关。因此，考虑船舶规格和环境条件参数的经验公式可以预测沿特定路线的航行时间。然后，对应于最短预测通过时间的路线成为最佳路线。在新西伯利亚群岛的周围应用了多元线性回归和LASSO回归，其中有3条路线。虽然获得的函数都是过拟合模型，不适合实际预测，但是LASSO可以估计更合理的标准化速度范围。分别通过多线性模型和LASSO模型预测的标准化速度范围介于每天310到1640 km和310到602 km。可以期望进行LASSO回归分析来确定在NSR上需要采取所有谨慎措施的地方。