一种用于多目标约束优化的改进进化算法

当前求解多目标优化的进化算法主要考虑如何处理相互冲突的多个目标间的优化,很少考虑对约束条件处理的问题.对此,给出了一种基于双群体搜索机制的改进差分进化算法,以求解多目标约束优化问题.采用两个不同种群,分别保存可行个体与不可行个体的双群体约束处理策略,利用基于Pareto的分类排序多目标优化技术,完成对进化个体解的评价.并通过群体混沌初始化、自适应交叉和变异操作来提高基本差分进化算法的性能.对三个经典测试函数的仿真结果表明,文中算法在均匀性、逼近性及收敛速度三方面均优于非支配排序遗传算法,而收敛速度也优于另两种改进进化算法.     

基于粒子群算法的冷连轧机轧制负荷分配优化

利用粒子群算法设计了一种冷连轧轧制负荷分配的优化方法。根据某1450五机架冷连轧机生产工况,以压下量分配为自变量,以轧制力成比例分配为目标函数,将压下量分配的约束条件作为惩罚项,建立惩罚函数。通过粒子群多代运算,求出罚函数值最小点,得到压下负荷最佳分配点。实验证明,粒子群算法在轧制负荷分配计算中,具有算法实现简便、运算速度快、收敛性好等优点,可以作为一种冷连轧轧制负荷分配的新方法加以推广。     

基于多目标种群协同算法的板坯入库优化

针对钢铁企业板坯入库决策问题,基于出库次序A型约束、分散性约束和垛位限高约束等构建了以板坯综合匹配度、垛位利用度和库存均衡度为目标函数的多目标入库决策优化模型。提出一种多目标种群协同粒子群优化算法,并设计了局部搜索策略以提高外部归档集中Pareto解的多样性,同时利用Pareto最优解改进粒子速度更新方式,达到多种群协同优化的目的。仿真实验证明,该算法可以更好地解决多目标板坯入库优化问题。     

基于遗传算法的轧制负荷分配优化

利用遗传算法设计了一种冷连轧轧制负荷分配的优化方法.根据某厂单机架可逆式冷轧机生产工况,以轧制力成比例分配为目标函数进行优化.实验证明,遗传算法在轧制负荷分配计算中,具有算法实现简便、运算速度快等优点,使用该策略能够获得良好的板形及生产的平稳过渡,可用于指导冷连轧生产.     

微分进化多目标优化算法研究

对现有的微分进化多目标优化算法做了进一步的研究,揭示了其在机械工程领域中应用存在的问题。在此基础上,对现有的微分进化多目标优化算法进行了修正,给出了基于微分进化算法、适用于机械工程领域的约束多目标优化算法。与目前普遍采用的粒子群多目标优化算法进行了比较,并通过颇具典型意义的双圆弧齿轮传动约束多目标优化范例进行了验证。研究结果表明,该方法比粒子群多目标优化算法具有更好的非劣解的多样性和收敛性,快得多的收敛速度,且程序设计简单、易懂。为机械工程领域约束多目标优化设计提供了一种切实可行的设计算法。     

基于过滤器技术的约束粒子群优化算法

工程设计中处理约束优化常采用罚函数法,但其优化结果敏感于惩罚因子,针对特定的实际问题往往需要多次试验以得到合适的罚因子取值。为了避免反复的参数选取测试过程,将过滤器约束处理机制和粒子群优化(Particle swarm optimization,PSO)相结合用于求解约束优化问题。过滤器方法基于多目标规划中的支配思想,以一组互不支配点所对应的目标值与违背度对构成过滤器,利用其处理约束可以避免使用罚函数。基于过滤器的约束PSO算法在粒子进化过程中,对各粒子历史最优解和粒子群历史最优解分别构造滤器,并依据可行性优先的粒子比较准则从对应的过滤器中选择最优解从而实现粒子的更新。然后,利用工程优化设计标准算例和翼型优化设计实例,将过滤器PSO算法和罚函数PSO算法、遗传算法进行比较研究,结果表明过滤器PSO算法能够获得较好的约束优化设计结果,是求解约束优化问题的一种有效方法。     

球面四杆机构函数综合的带修复策略可行性规则差分进化算法

根据杆长限制条件,建立约束方程,进而得出求解球面四杆机构函数综合问题的非线性方程组,并将该方程组的求解转化为鞍点规划问题。以杆长协调、传动角、避免乱支缺陷等为约束条件,提出球面四杆机构近似函数综合的约束优化模型,再应用差分进化(Differential evolution,DE)算法求解该问题。在定义约束违反度和弱、强不可行解的基础上,提出处理约束条件的改进可行性规则,形成求解约束优化问题的可行性规则差分进化(Feasibility-rule-based DE,FRDE)算法。应用4个benchmark约束优化问题测试FRDE算法的优化性能,结果表明,其可靠性和稳健性指标优于对比算法。面向机构优化综合问题,将修复策略融入FRDE算法,发展为带修复策略的FRDE算法(Feasibility-rule-based DE algorithm with repair strategies,FRRDE)。给出5个函数综合实例。结果显示,优化模型和方法可行有效,且FRRDE算法的优化性能好于对比算法。     

基于惩罚和修复策略的约束优化遗传算法

约束优化问题中最难以解决的就是约束处理问题,将惩罚函数法与修复策略相结合应用于非线性约束优化遗传算法之中,使得约束优化问题在惩罚函数和修复算子的协同作用下收敛于全局最优,有效避免了迭代过程中大量非可行解的产生,解决了在遗传算法约束优化问题中单独使用惩罚和修复方法时一些难以解决的问题。基于随机方向法构造的修复算子作用效果显著,采用多个测试函数对算法进行检验,均能较好地收敛于可行域中的最优解,验证了算法的可靠性。     

一种多目标多约束优化设计的高效包络函数法

实际工程结构优化问题往往具有多个约束甚至多个目标,约束数目与目标数目的缩减对提高优化计算效率具有重要意义。利用包络函数可以使多约束、多目标优化问题转化为单目标、单约束优化问题,甚至可以使具有多个约束和多个目标的优化问题转化为具有单目标的无约束优化问题。提出k次方根包络函数,这种包络函数具有表达简捷、意义明确、方便易用的优点。大量结构优化应用实例证实,这种包络函数的有效性和优越性。     

圆柱齿轮传动多目标优化的差分进化算法

针对单级斜齿圆柱齿轮传动机构优化设计问题,建立以体积最小化和重合度最大化为目标的约束多目标优化模型。为提高Pareto前沿的分布均匀性和分布广度,将网格Pareto占优技术与约束多目标差分进化算法结合,设计网格占优约束多目标差分进化算法(ε-CMODE)。根据工程实践需要,将离散约束多目标优化模型映射为约束非负整数规划问题,再改进ε-CMODE算法以求解该模型。最后,给出优化设计实例。结果表明,ε-CMODE算法能有效求解齿轮机构多目标优化问题,得到均匀分布的Pareto前沿,可为设计人员提供多组备选解。     

基于改进多目标粒子群算法的冷连轧规程优化设计

合理的轧制规程能够提高轧机的产量和产品的质量,带来显著的经济效益。采用多目标粒子群算法,选择等相对负荷和预防打滑为目标进行冷连轧规程优化。针对算法存在的收敛性和分布性难以均衡的问题,引入一种基于平行坐标系的密度和收敛潜能计算方法;同时,为克服算法易于陷入局部最优的缺陷,提出一种带个体扰动的全局最优领导粒子选择策略。仿真结果表明,该方法能快速跳出局部极值,获得具有更好收敛性和分布性的近似Pareto前沿。最后应用该方法对某五机架冷连轧机进行了轧制规程优化。       

基于遗传算法的冷连轧轧制规程优化设计

合理的轧制规程是使轧制过程达到最佳状态的重要保证。轧制规程涉及的有关数学模型和需要考虑的约束条件较多，因此其优化设计也较为复杂。应用遗传算法进行轧制规程优化设计，即以负荷均衡和板形良好为目标，在一定的约束条件下，对压下率分配比进行编码。通过遗传算子操作并保留每代优势个体，从而得到冷连轧轧制规程。对某1370冷连轧机进行设计比较，表明该方法优化速度快，适合在实际中应用。     

NEW MULTIOBJECTIVE SIMULATED ANNEALING ALGORITHM AND ITS APPLICATION TO TURBINE CASCADES’ DESIGN

０ＩＮＴＲＯＤＵＣＴＩＯＮＳｉｍｕｌａｔｅｄａｎｎｅａｌｉｎｇａｌｇｏｒｉｔｈｍｉｓａｋｉｎｄｏｆｅｆｉｃｉｅｎｔａｌｇｏｒｉｔｈｍｐｒｏｐｏｓｅｄｉｎｒｅｃｅｎｔｙｅａｒｓｆｏｒｓｏｌｖｉｎｇｌａｒｇｅｓｃａｌａｒｃｏｍｂｉｎａｔｏｒｉａｌｏｐｔ...     

An integrated solution—KAGFM for mass customization in customer-oriented product design under cloud manufacturing environment

In hot strip rolling process, rolling schedule is a key technology which directly influences strip product quality. Rolling schedule optimization is actually a problem of load distribution. To make a better rule of the load distribution of aluminum hot tandem rolling, multi-objective optimization algorithm is used to optimize rolling schedule. Preventing slipping, power margin and minimum energy consumption are selected as the optimization objectives. To make a precision calculation of rolling schedule, an adaptive neural network which is based on classification system is applied to improve the prediction ability for the rolling force, and its on-line training system reduces the prediction errors caused by different rolling conditions. The improved differential evolution algorithm is used to search the Pareto front, and it obtains a good approximation of the Pareto-front and decreases computation time. Load distribution strategies focused on different objectives are generated from the Pareto front to meet the requirements of industrial spots. The experiment result shows the algorithm covers the front quickly and distributes well. Comparing with the original schedule, the proposed method reduces the probability of slippage and energy consumption.     

NEW MULTIOBJECTIVE SIMULATED ANNEALING ALGORITHM AND ITS APPLICATION TO TURBINE CASCADES' DESIGN

A new kind of multiobjective simulated annealing algorithm is proposed,in which the concept of non-dominate d character is introduced and a new multiobjective acceptance criterion is set u p.The optimization example of a typical mathematical problem with two minimum ob jective functions indicates that all of the solutions contract to the set of the non-dominated points,and the variation trend of the optimal solutions is verif ied to be identical with that obtained using Genetic Algor thms.The ne w developed algorithm is then applied to the multiobjective optimization design of turbine cascades,in which it is coupled with the aerodynamics computation of the cascade flow fields and performance and the calculated loss coefficient and work potential of the cascade are considered as the objective functions,thus set ting up a technique to the engineering optimization design for the cascades.The optimization results,by the view of a group of optimal solutions,show that the a lgorithm is superior to the traditional technique of multiobjective optimization design and can be applied to more than two objective optimization cascade desig n problem or other engineering multiobjective optimization designs.     

Multiobjective trajectory optimization of intelligent electro-hydraulic shovel

Multiobjective trajectory planning is still face challenges due to certain practical requirements and multiple contradicting objectives optimized simultaneously. In this paper, a multiobjective trajectory optimization approach that sets energy consumption, execution time, and excavation volume as the objective functions is presented for the electro-hydraulic shovel (EHS). The proposed cubic polynomial S-curve is employed to plan the crowd and hoist speed of EHS. Then, a novel hybrid constrained multiobjective evolutionary algorithm based on decomposition is proposed to deal with this constrained multiobjective optimization problem. The normalization of objectives is introduced to minimize the unfavorable effect of orders of magnitude. A novel hybrid constraint handling approach based on ε-constraint and the adaptive penalty function method is utilized to discover infeasible solution information and improve population diversity. Finally, the entropy weight technique for order preference by similarity to an ideal solution method is used to select the most satisfied solution from the Pareto optimal set. The performance of the proposed strategy is validated and analyzed by a series of simulation and experimental studies. Results show that the proposed approach can provide the high-quality Pareto optimal solutions and outperforms other trajectory optimization schemes investigated in this article.     

Multidimensional visualization and clustering for multiobjective optimization of artificial satellite heat pipe design

This study presents a newly developed approach for visualization of Pareto and quasi-Pareto solutions of a multiobjective design problem for the heat piping system in an artificial satellite. Given conflicting objective functions, multiobjective optimization requires both a search algorithm to find optimal solutions and a decision-making process for finalizing a design solution. This type of multiobjective optimization problem may easily induce equally optimized multple solutions such as Pareto solutions, quasi-Pareto solutions, and feasible solutions. Here, a multidimensional visualization and clustering technique is used for visualization of Pareto solutions. The proposed approach can support engineering decisions in the design of the heat piping system in artificial satellites. Design considerations for heat piping system need to simultaneously satisfy dual conditions such as thermal robustness and overall limitation of the total weight of the system. The proposed visualization and clustering technique can be a valuable design tool for the heat piping system, in which reliable decision-making has been frequently hindered by the conflicting nature of objective functions in conventional approaches.     

K-S函数在多目标优化中的应用

提出了一种将约束优化问题转化成无约束优化问题的新技术 ,并在此基础上提出了一种新的多目标优化算法 K- S法。该算法通过 Kreisselm erier- Steinhauser函数构造出标准化目标函数、行为约束函数的包线 ,通过求包线的无约束极值而求解出原多目标优化问题的解。该蒜法无需取加权因子 ,也无需对种目标函数进行分离优化 ,并且设计初始点选择不受限制。本文最后给出了一个典型的多目标优化算例 ,结果令人满意     

冷轧超高强钢平直度与断面形状前馈控制技术

针对六机架冷连轧机组超高强钢生产过程中因机架数多、控制工艺复杂、各机架控制能力得不到充分发挥而导致成品平直度与断面形状超差的问题,充分结合六机架冷连轧机组的设备与工艺特点,基于超高强钢冷轧过程平直度与断面形状预报模型,分析了超高强钢冷连轧过程平直度与断面形状前馈控制策略,提出了基于来料的各机架出口目标平直度曲线设定方法,开发了适合于六机架冷连轧机组的超高强钢平直度与断面形状前馈控制技术。将该技术应用到某钢厂六机架冷连轧机组,各机架控制能力得到了充分发挥,成品平直度比该技术应用前提高了12.5%,且断面形状分布更加均匀,具有进一步推广应用的价值。