农机跨区作业紧急调配算法适宜性选择

目前农机跨区紧急作业中供需信息不对称,农机部门缺乏科学合理的紧急调配方案,无法在紧急状况下指导农机进行及时有效的调配。针对上述问题,该文研究了农机跨区作业紧急调配模型和算法。首先分析了多机多任务紧急调配过程,建立了以最小化调配成本和损失为目标的紧急调配模型,提出了基于距离最近优先的多机多任务紧急调配算法(shortest-distance first algorithm,SDFA)和基于贡献度最大优先的多机多任务紧急调配算法(max-ability first algorithm,MAFA),前者是搜索当前距离最近的农田和农机,进行优先分配,后者是搜索当前贡献度最大的农田和农机并进行优先分配。采用该文算法对河北省邯郸市2017年的真实数据以及随机生成的农田和农机实例库进行计算与分析可知,当农机数量充足时,算法MAFA的平均调配成本要比SDFA的平均调配成本降低4.34%。当农机不足时,SDFA的平均损失和平均调配成本要比MAFA的平均损失和平均调配成本分别下降了12.79%和4.11%。进一步验证可知,当农田数量为6时,上述2种算法比笔者之前提出的基于非合作博弈紧急调配算法(non-cooperative game algorithm,NCGA)的平均运算性能均提升25%以上,当农田数量为30时,性能均提升41%。该研究可为农机管理部门紧急调配与决策分析提供科学依据。

Suitability selection of emergency scheduling and allocating algorithm of agricultural machinery

Abstract: At present, the supply and demand information in the emergency operation of agricultural machinery is asymmetric. Without the scientific and reasonable emergency allocation plan, agricultural machinery department could not guide scheduling and allocating of agricultural machinery timely and effectively in the emergency situation. To solve the above problems, models and algorithms of emergency scheduling and allocating based on intelligent agricultural machinery platform are studied in the paper. The intelligent agricultural machinery platform, which is integrated of GPS (global positioning system), information communications, networking, Internet of things and other related technologies, can realize real-time data collection, data transmission, data storage, data calculation and decision-making of agricultural machinery and farmland. Emergency scheduling and allocating problem based on intelligent agricultural machinery platform needs to establish one kind of mapping relation between farmlands and agricultural machinery. The main influencing factors including time limit, location, distance and operation ability in emergency deployment are analyzed and the mathematical model of the emergency scheduling and allocating problem is established with minimizing costs and losses as scheduling and allocating objectives. According to the model, the emergency scheduling and allocating algorithm based on the shortest-distance first algorithm (SDFA) and the emergency scheduling and allocating algorithm based on the max-ability first algorithm (MAFA) are proposed in this paper. The former one is prior to allocate the agricultural machinery to the nearest farmland and the latter one is prior to allocate the agricultural machinery with the maximum operation ability. Taking the emergency operations of wheat combine harvesters in Handan City, Hebei Province as the research objects, the emergency scheduling and allocating scheme is calculated by using the algorithms proposed in the paper. The results show that the cost of MAFA algorithm is lower than that of SDFA algorithm when the amount of agricultural machinery is sufficient, and therefore the MAFA algorithm is more suitable for the situation with enough agricultural machinery. Costs and losses of SDFA algorithm are lower than those of MAFA algorithm when the amount of agricultural machinery is insufficient, so SDFA algorithm is more suitable for the situation with the shortage of agricultural machinery. Further analysis shows that when the amount of farmland is 6, the average running time of the emergency scheduling and allocating algorithm with non-cooperative game algorithm (NCGA) proposed in the existing literature is 3.215 s, the average running time of both MAFA and SDFA is less than 2.4 s, and the performance has been improved by more than 25%. When the amount of farmland is 10, the average running time of NCGA is 4.286 s, the average running time of both MAFA and SDFA is less than 2.7 s, and the performance has been improved by more than 37%. When the amount of farmland is 15, the average running time of NCGA is 5.369 s, the average running time of both MAFA and SDFA is less than 3.3 s, and the performance has been improved by more than 39%. When the amount of farmland is 30, the average running time of NCGA is 6.485 s, the average running time of both MAFA and SDFA is less than 3.8 s, and the performance has been improved by more than 41%. The experimental results show that the 2 emergency scheduling algorithms have better operational performance and can provide scientific basis to the emergency allocation and decision analysis for agricultural machinery management department.