温室穴盘苗自动移钵路径优化

为了优化移栽机补栽作业时的移钵路径,该研究基于免疫算法构建了克隆选择算法和免疫遗传算法2种适合求解移钵路径优化问题的模型,并与固定顺序法和遗传算法对比,进行移钵路径规划仿真试验和验证试验。结果表明:克隆选择算法模型和免疫遗传算法模型均能有效优化移钵路径,免疫遗传算法模型的路径规划效率较高,而克隆选择算法模型的路径规划效率较低。验证试验条件下,该研究2种模型的路径规划长度分别为48 977和48 945 mm,相比固定顺序法分别缩短7.59%和7.65%,相比遗传算法模型分别缩短3.60%和3.66%;2种模型的计算时间分别为5.86和2.72s,免疫遗传算法模型的计算时间相对遗传算法减少15.79%。免疫遗传算法模型可作为温室穴盘苗后续机械化批量补栽的路径规划控制基础。

Optimization of automatic transplanting path for plug seedlings in greenhouse

In mechanized plug seedling production, it is necessary to eliminate inferior plug seedlings and replant healthy plug seedlings with transplanter. In order to improve the plug seedling transplanting efficiency, plug seedling transplanting path would be planned. In this paper, the Clone Selection Algorithm(CSA) model and Immune Genetic Algorithm(IGA) model were constructed to solve the problem of transplanting path optimization of plug seedlings in greenhouse. Compared with the fixed sequence method and the Genetic Algorithm(GA), the simulation and verification tests of the transplanting path planning were carried out with 50-hole, 72-hole, and 128-hole plug trays. Set the population size in the CSA model to 40, the mutation probability to 0.4, and the number of clones to 10. Set the total population size in the IGA model to 40 (of which the memory cell bank capacity was 10), and the crossover probability to 0.5, the mutation probability was 0.4, and the diversity evaluation parameter was 0.95. The results showed that CSA and IGA model designed in this paper could achieve the purpose of optimizing the plug seedling transplanting path. The planned path length of the two models was similar, and the planned path length of two models was significantly shortened compared with the common sequence method and the GA. In the verification tests, the planned path length of common sequence method, GA model, CSA model and IGA model were 52 998, 50 807, 48 977 and 48 945 mm respectively. Compared with the common sequence method and the GA, the planned path length of CSA model and IGA model were shortened by 7.59% and 7.65%, and shortened by 3.60% and 3.66% compared with the GA model. The path planning efficiency of IGA model was higher than that of the CSA model, the calculation time of GA model, CSA model and IGA model was 3.23, 5.86 and 2.72 s respectively, the calculation time of IGA model was 15.79% less than that of the GA model. The models designed in this paper were suitable for path optimization of various-size plug seedlings automatic transplanting in greenhouse, it only needed to get the coordinates of plug seedlings according to the actual needs.