舰艇编队海上运输补给物资需求预测方法

Predictive method of supply demand for sea transportation and replenishment of naval ship group

按照是否与任务、事件相关,将物资需求分为两类,对于任务、事件相关物资需求的预测,将任务进行合理的分解,并根据物资消耗与任务、事件之间的关系,给出了预测的一般模型;对于与任务、事件联系不紧密的物资需求的预测,则根据历史经验及该物资固有的消耗规律,提出了经验预测模型。为了解决舰艇编队海上运输补给物资需求预测所存在的问题,利用案例推理的方法生成了预测所需的样本数据,以最小二乘向量机(LSSVM)模型为预测模型,并以岛屿进攻作战的防空弹药需求预测为例进行了实例分析。结果表明:案例推理生成的样本数据可用,选用LSSVM模型的预测结果与其他预测模型表现出了一致性,但LSSVM相对误差较小;该方法在某种程度上解决了样本数据有限的问题,适用于作战物资需求的预测问题。

According to the supply demand being related to events or not,the demand can be divided into two kinds.For the supplies relative to events,the mission should be decomposed reasonably.Then,ageneral forecasting model is presented according to the relationship between the supply consumption and the events.For the supplies irrelative to events,ageneral model is proposed according to the historical experience and the inherent characteristic of supply consumption.To solve the problems of demand forecasting,the case-based reasoning is used to create samples.The LSSVM model is used as a forecasting model,and the anti-air ammo demand forecasting for an island offensive operation is taken as an empirical analysis.The result indicates that the samples created by CBR are available,and that the forecasting results of different models are consilient but more accurate.This method can solve the problem of the lack of samples,so it is applicable to the predication of combat demand for supplies.