基于遗传算法的装载机工作装置优化设计应用研究

【作者】 熊雪峰；

【导师】 孙骊； 郭康权；

【作者基本信息】 西北农林科技大学 ， 机械设计与理论， 2001， 硕士

【摘要】 本文根据目前工程机械领域中的高效率、低能耗的发展方向，运用一种智能随机优化方法——实值编码遗传算法，来优化装载机的工作装置。目的在于解决运用传统的优化方法设计工作装置时，不容易得到全局最优结果的弊端。 本研究从装载机工作装置运动分析入手，在满足机构性能要求的前提下，建立了以动臂油缸和翻转油缸处于铲掘位置时的最大举升力和最大铲掘力为目标函数，以连杆机构的各个杆件长度为设计变量，以工作装置的工作性能和结构设计要求为约束函数的多维优化数学模型。根据本优化设计的实际情况，本研究对标准遗传算法进行了改进：将标准遗传算法和混合惩罚函数的思想相结合，设计了合理的适应度函数；采用了实值编码的编码方式，从而避免了进行二进制编码时由于染色体的位串过长导致的重复操作、降低杂交和变异的效果、增加运算的时间和陷于局部最优的可能，减少了标准遗传算法中译码的步骤；采用了联赛选择机制、算术交叉、非均匀变异三种遗传算子。所有的程序都用TurbC2.0编制而成。该程序由选择、交叉变异、群体更新等主要模块及输入、输出、编码、适应度计算等辅助模块构成。 通过具体实例分析，进行了改进遗传算法的参数优化配置，选择出了合理的最大群体规模、最大遗传代数、选择概率、交叉变异概率、非均匀度变异参数等。结果表明装载机在满足最大卸载高度、最小铲掘深度、最大高度时的卸载距离、平移特性、卸料性、自动放平性等技术要求的前提下，其动力性能比原来提高了大约9％。优化设计所用的时间大大缩短，执行效率是混合惩罚函数的3倍。 改进遗传算法成功的解决了装载机工作装置优化设计中由于目标函数非线性、自变量维数多、约束函数多从而求解困难的问题，提高了工程设计水平和设计效率，缩短了设计周期，减少了工程投资。更多还原

【Abstract】 An Application Research of the Work-equipmentOptimization of Loder-dozer Based On Genetic AlgorithmGraduate:Xiong XuefengTutor:Prof. Sun LiProf. Guo KangquanAbstractAccording to the direction of high efficiency and saving-energy in engineering machinery, this paper bring forward a kind of intelligence randomizing optimizing method桮enetic Algorithm (GA), to optimize the equipment of the loader-dozer. The purpose of this paper is to solve the problem of not easy to derive a global optimal result when using traditional optimizing method.Tiis paper make a movement analysis of work equipment, establish a high-dimensional optimal mathematical model; The objective functions of this model are the lifting force of arm working cylinder and the shoveling force of turn over working cylinder, the design variable is the length of each member of link gear, the constraint function is the design requirement. According to the factual conditions of this design, The paper improved the traditional Genetic Algorithm by several ways: Unites the penalty thoughts and standard GA, and designs a suitable fitness function; Adopts a improved encoding??real-value encoding, Because binary encoding may result reoperation, lowing efficiency of crossing and mutation, and increasing the run-time and the possibility of local optimal due to the long string of chromosome. This encoding method avoids all these problem successfully, and tail off the decoding step in GA; Adopts tournament selection, mathematical crossover, non-linear mutation. All of programs compiled with Tc2.O. The programs are made up of main model such as selection, crossover, mutation and generation and so on and supplementary model such as input, output, encoding and fitness calculating and so on.Through sample analysis, this paper takes parameter optimization of improved GA, and selected reasonable parameters such as: popsize, maxgen, Pr, Pr, Pm, non-linear mutation parameter. The results shows that the dynamic force performance of loader-dozer is up by 9% under the premises of filling the specification such as maximum discharge height, minimum shovel depth, discharge distance, translation characteristic, unloading characteristic, auto laying characteristic and so on. The optimization time is shorten greatly, perform efficiency of improved GA is 3 times of the penalty.Improved GA successfully solved the problem of difficult-solution due to the nonlinear, multibounding, high dimensional of objective function in working equipment optimization of loader-dozer, increased the efficiency and level of engineering design, and shorten the periodic of design, decreased the capital.更多还原