应用层次分析法计算分插机构优化目标的权重

Using AHP to calculate optimization objective weights of transplanting mechanism

分插机构优化是一个多目标、多参数优化问题。该文在分插机构可视化辅助分析优化平台上得到多组可行优化结果之后,为进一步评价出最优解,对评判方法中权值的求解方法进行了应用分析。将分插机构的12个优化目标中的10个作为评判方法的评价指标,分别应用层次分析法中的0～2三标度法和1～9标度法确定各评价指标的权值,即通过建立初始评判矩阵和比较矩阵,并通过一致性检验得到最终判断矩阵,继而得到所有目标的权重集,其结果分别为(0.6216,0.2450,0.0881,0.0302,0.0101,0.0033,0.0011,0.0004,0.0001,0.0001)和(0.6892,0.2233,0.0620,0.0175,0.0052,0.0017,0.0006,0.0003,0.0001,0.0001)。2组结果均与实际相符,1～9标度法结果精度更高。试验结果表明,应用AHP求解分插机构的优化目标权值是合理、客观、有效的。此方法可为一般农机多目标优化问题的权值求解起到一定的借鉴作用。

Transplanting mechanism optimization is a strong coupling, fuzziness and nonlinearity optimization problem which involves multi-objective and multi-parameter, and the optimization results are pareto. For these feasible solutions, how to look for the best optimization result is the key issue to improve the transplanting mechanism optimization design. At present, only some agricultural experts can identify which one is the best through the motion trajectory shape. But sometimes which has the subtle difference between some results, and the experts cannot judge which one is more optimal. Based on the pareto achieved on the computer-assisted analysis software of the transplanting mechanism, for further evaluating the best optimal solution, the weights evaluation method of the calculation method was analyzed. There are twelve optimization goals for the transplanting mechanism on the high-speed-type rice transplanter. They are given as following： 1） When the transplanting mechanism is operating, the two transplanting arms do not interfere each other; 2） When transplanting arms fetch seedlings, the supporting part of the seedling needle do not interfere with seedling box; 3） The fetched seedling block is oblong, and the trace of fetching seedlings at the seedling box is vertical; 4） The transplanting arms shaft should not contact with the lower part of the transplanted seedlings; 5） The angular difference between the angles of pushing and fetching seedlings should be between 55°and 60°; 6） The distance between seedling-separating needle tip and seedling-pushing needle tip is more than 260 mm; 7） The opening length along the direction of transplanter travel, formed by the absolute motion trajectory must be less than 30 mm; 8） The wrap angle between the seedling needle and horizontal should remain 65°to 85°when pushing seedlings; 9） To ensure the uprightness of the seedlings after transplanting, the wrap angle between the seedling needle and horizontal line should remain-10°to 20°when fetching seedlings; 10） The planetary gear shaft should not contact with the lower part of the transplanted seedlings; 11） The gear modulus is more than 2.5 mm; 12） The distance between disk bottom and ground is more than 20 mm. In the above goals, the first two goals are the movement interference judgment of the transplanter, and can be get with accurate decision results. The rest ten goals are certain fuzziness, namely the target decomposition to each goal function value is not the only value but a range, all values of the range can satisfy the kinematic movement requirement. Therefore the rest ten optimization goals are chosen as the evaluation index of the evaluation method in this article, the analytic hierarchy process in 0~2 three-scale method and 1~9 scaling method are applied to determine the weight of each evaluation index, by establishing the initial judgment matrix and comparison matrix. The final judgment matrix is got through the consistency test, and then all the target weights set is got, which is （0.6216, 0.2450, 0.0881, 0.0302, 0.0101, 0.0033, 0.0011, 0.0004, 0.0001, 0.0001） and （0.6892, 0.2233, 0.0620, 0.0175, 0.0052, 0.0017, 0.0006, 0.0003, 0.0001, 0.0001） , respectively. These two kinds of results are consistent with the reality, and the precision of the 1~9 scale results is higher. The results show that using the AHP to solve the weights of the transplanting mechanism optimization target is rational, objective and effective. This method can be a reference for the general multi-objective optimization problem to solve the weights.