一类带有一个反凸约束全局规划的收敛算法

讨论一类带有一个反凸约束的全局规划（Ｐ）：ｇｌｏｂａｌｍｉｎｆ（ｘ）＝ＣＴｘ，ｓ．ｔ．ｘ∈Ｄ＝｛ｘ｜ｈ（ｘ）≤０｝和ｇ（ｘ）≥ｏ，其中Ｃ≤Ｒｎ，ｇ（ｘ），ｈ（ｘ）是Ｒｎ上的有限凸函数。我们给出这类问题的一个外切型算法。在不需要稳定性假定的一般意义下，证明了算法有限终止于（Ｐ）的全局解，或者算法产生一个收敛到全局解的点列。     

基于AEA和PSO的双层同步换热网络综合方法研究

换热网络优化是一个多维、非凸、非线性、不连续的复杂混合整数非线性规划问题,传统的优化方法很难寻找到全局最优解。针对该问题研究建立了换热网络分级超结构非等温混合模型,提出了一种双层同步优化算法。该算法外层使用Alopex Evaluation Algorithm(AEA)算法优化结构变量分流比,内层使用Particle Swarm Optimizer(PSO)算法优化换热量。还提出了一种改进的不可行解修复策略,改善了算法的搜索能力。三个案例研究用以说明算法可以稳定寻找到较好的优化结果。     

考虑再生循环的多杂质用水网络全局优化

引入再生循环单元可以有效地减少多杂质用水系统整体的新鲜水消耗量和废水排放量。其现有的数学规划模型是一个非凸性严重的混合整数非线性规划问题,直接求解不仅计算难度大而且求解速度慢,这样只能得到局部最优解。本文利用分段线性松弛的方法对该模型进行处理,将双线性等式约束转换成线性约束,从而使复杂的水网络模型转换为对应的混合整数线性规划问题,因现有的商业优化软件对混合整数线性规划问题的求解算法已经非常成熟,这样降低了计算难度并取得全局最优解。算例结果表明分段线性松弛的方法可以使求解简单,能够获得全局最优解。管线费用的减少使年度化费用降低,全局优化能得到一个经济性更好的多杂质水网络,这种方法对于节省设备投资和操作费用,具有重要意义。     

带有反凸约束凸规划的锥分解算法

将凹规划问题的锥剖分方法应用于带有一个反凸约束的凸规则，得到了一个锥分解算法，并证明了算法的收敛性。     

基于群体智能算法的换热网络同步最优综合

换热网络同步综合方法一般需要建立复杂的混合整数非线性数学规划模型,该模型具有非凸、非线、不连续的特点,属于最难求解的一类NP-hard问题,应用传统的优化算法很难确定其全局最优解,尤其是对大规模换热网络综合问题,甚至无法在合理时间内接近全局最优的局部最优解。针对换热网络同步综合问题,提出基于群体智能算法的分层优化策略,外层采用离散粒子群算法与遗传算法相结合的混合群体智能算法优化换热网络结构,内层在结构变量给定条件下利用改进粒子群算法优化冷热物流分流比与换热负荷。两个典型算例研究证明了该方法能以较高的效率和稳定性得到较好的优化结果。     

随机需求VRP的一个算法

研究了随机需求VRP问题,在服务仅能路由失败一次和不允许部分服务的情况下,给出了两阶段Clarke-Wright模拟退火算法。对需求为均匀分布的VRP问题做了数值试验,给出了具体的方案。数值结果表明这个算法比简单的两阶段模拟退火算法好,而且算法简单,容易实现。这个算法在第Ⅱ阶段模拟退火算法开始时给出了一个比较合理的初始解,较好地处理了大范围交叉的问题,从而使得模拟退火算法快速收敛到近似最优解。     

基于CSTR的反应器网络综合双层优化算法

反应器网络综合问题一般都是复杂的非线性规划问题,在分析基于全混流反应器的反应器网络模型特点的基础上,提出了求解该模型的双层优化算法. 通过将反应器网络综合非线性规划问题分解为物流流量和反应器体积空间的线性优化和浓度空间的优化搜索问题,降低了所求解问题的规模和难度,同时利用全局优化算法进行浓度空间的优化搜索,提高了求得全局最优解的概率. 实例研究表明,双层优化算法可以更准确地给出最优的反应器网络结构以及网络中反应器的类型和大小.     

基于涡流搜索算法的支持向量机分类模型

提出一种将涡流搜索算法用于支持向量机参数选取的新算法,利用该算法不必遍历搜索空间内所有的参数点即可找到全局最优解。给出了具体的算法流程,并进行了仿真。仿真实验结果表明涡流搜索算法是选取SVM参数的有效方法。     

典型石化工艺装置换热网络模型的全局优化策略

在石化节能的工程实践中,由于冷物流分流限制,石化工艺装置中存在一类单一冷物流被多条热物流加热的问题。针对这类典型的石化工艺过程换热流程优化模型存在陷入局部最优解或无解的困境,提出利用凸化非凸项技术,将非线性规划问题转化为凸规划问题。通过逐段线性化的方法松弛原有可行区间,并依据上下界差异确定分段数值,最终得到目标函数的全局最优解。以延迟焦化装置的换热网络为例,计算结果显示采用全局优化策略计算的结果更优,减少约3.4×10~6 CNY×a~(-1)。     

基于改进蚁群算法的PCB孔加工路径优化

针对于印制电路板(PCB)孔加工过程中刀具的路径规划问题,提出了一种改进的蚁群算法,使信息素浓度可以更好地反映路径信息,解决了经典蚁群算法易收敛到局部最优解的问题。仿真实验结果表明:改进算法不仅增强了对于解空间的搜索能力,而且提高了搜索到全局最优解的概率。     

Prediction‐correction method for optimization of simulated moving bed chromatography

A systematic algorithm for simulated moving bed (SMB) chromatography process development that utilizes dynamic optimization, transient experimental data, and parameter estimation to arrive at optimal operating conditions is described. These operating conditions ensure both high purity constraints and optimal productivity are satisfied. This algorithm proceeds until the SMB process is optimized without manual tuning. In a case study, it has been shown with a linear isotherm system that the optimal operating conditions can be reached in only two changes of operating conditions following the proposed algorithm. Another case study with a linear isotherm system has shown that the algorithm is robust to optimize the SMB even if there is significant model mismatch at first. © 2012 American Institute of Chemical Engineers AIChE J, 59: 736–746, 2013     

分散常规控制系统结构设计的遗传算法求解

化工过程通常为高维大系统,在进行底层常规PID控制系统设计时,难以得到最佳配对方案,迫切需要一种简便且快速的计算机优化算法解决该问题。对此提出了一种基于遗传算法的分散常规控制系统结构设计的优化求解算法。该算法基于遗传算法的思想,对结构矩阵进行字符串形式的编码,将变量配对转换成数组形式,选择合适的遗传算子,通过交叉、变异、选择等遗传操作,解得无约束条件下的最优变量配对方案。其次,结合约束条件对遗传算子适当改进和调整,以无约束条件下的最优配对方案为基础,仅需通过变异和选择算子,就可解得存在约束条件的最优变量配对方案,并且具有较快的计算速度。最后通过包括TE过程在内的实例分析证明了该算法的有效性与准确性。     

DOBD Algorithm for Training Neural Network: Part II. Application

In the first part of the article, a new algorithm for pruning networkDynamic Optimal Brain Damage(DOBD) is introduced. In this part, two cases and an industrial application are worked out to test the new algorithm. It is verified that the algorithm can obtain good generalization through deleting weight parameters with low sensitivities dynamically and get better result than the Marquardt algorithm or the cross-validation method. Although the initial construction of network may be different, the finial number of free weights pruned by the DOBD algorithm is similar and the number is just close to the optimal number of free weights. The algorithm is also helpful to design the optimal structure of network.     

一种改进的混合遗传算法及在化工过程动态优化中的应用(英文)

The solutions of dynamic optimization problems are usually very difficult due to their highly nonlinear and multidimensional nature. Genetic algorithm (GA) has been proved to be a feasible method when the gradient is difficult to calculate. Its advantage is that the control profiles at all time stages are optimized simultaneously, but its convergence is very slow in the later period of evolution and it is easily trapped in the local optimum. In this study, a hybrid improved genetic algorithm (HIGA) for solving dynamic optimization problems is proposed to overcome these defects. Simplex method (SM) is used to perform the local search in the neighborhood of the optimal solution. By using SM, the ideal searching direction of global optimal solution could be found as soon as possible and the convergence speed of the algorithm is improved. The hybrid algorithm presents some improvements, such as protecting the best individual, accepting immigrations, as well as employing adaptive crossover and Gaussian mutation operators. The efficiency of the proposed algorithm is demonstrated by solving several dynamic optimization problems. At last, HIGA is applied to the optimal production of secreted protein in a fed batch reactor and the optimal feed-rate found by HIGA is effective and relatively stable.     

Selecting the Target Transformations in the Design of Chemical Production Systems

The design of production systems at chemical plants is considered. Attention focuses on the selection of the target transformations of materials within a virtual system. The information approach is adopted in an algorithm for step-by-step selection of the target processes and their optimal integration within the virtual system. The key feature of this algorithm is that it allows to use general principles for the distribution of the system’s functions between its elements and subsystems to ensure optimal integration of the target processes. Those principles were described in previous works. As an example, the algorithm is employed in the optimal organization of a system for converting lignite to synthesis gas, with specified proportions of the key components.     

应用于换热网络优化的并行双层RWCE算法

换热网络规模越大,其解空间内极值点呈指数性增长,优化时不仅要求算法具有强大的全局寻优能力,局部解空间的高精度搜索也不可或缺。鉴于强制进化随机游走算法(RWCE)优化换热网络时难以兼顾局部搜索能力、易导致遗漏最优解的现象,同时为增大算法优化大规模换热网络的种群数量,本文将精细搜索和并行计算相结合提出了并行双层RWCE算法。算法基于多核并行技术,通过并行线程分配建立基础层和精细层,基础层在并行计算技术加持下,算法全局搜索能力大幅提升,精细层将基础层当前最优解实时精细搜索,避免了原算法差解代替优解现象。最后通过两个算例进行验证,结果表明并行双层RWCE算法不仅具有更强的全局搜索能力,且兼具高精度的局部搜索能力,在优化进程中有效保护了最优解。     

受扰双线性系统的近似最优扰动抑制方法

研究具有外界持续扰动作用下双线性系统的最优控制问题.关于二次型性能指标给出了一种设计最优扰动抑制控制律的逐次逼近方法.利用该算法可将在扰动作用下双线性系统的最优控制问题转化为求解一组线性非齐次两点边值序列问题.通过迭代序列得到的最优扰动抑制控制律由解析的线性前馈-反馈项和序列极限形式的非线性补偿项组成.通过截取非线性补偿序列的有限项,可以得到近似最优扰动抑制控制律.仿真结果表明,该方法抑制外部持续扰动的鲁棒性优于经典反馈最优控制.     

Optimisation of fed-batch bioreactors using genetic algorithms

An optimisation procedure based on genetic algorithm approach is developed for the determination of substrate feed profiles for the optimal operation of fed-batch bioreactors. The problem specific knowledge generated through the rigorous application of the optimal control theory is used to formulate the set of decision variables representing the qualitative and quantitative aspects of the feed rate profile. A customized genetic algorithm with suitable genetic operators is used for generating the optimal feed profiles. Even though the optimal control theory is not explicitly used, the feed rate policies thus evolved are shown to retain the characteristics of the profiles generated through the application of optimal control theory. The efficiency of the proposed algorithm is demonstrated with two fermentation processes: secreted protein and yeast cell mass production.     

ADAPTIVE OPTIMIZATION OF CONTINUOUS BIOREACTOR USING NEURAL NETWORK MODEL

An adaptive optimization algorithm using backpropogation neural network model for dynamic identification is developed. The algorithm is applied to maximize the cellular productivity of a continuous culture of baker's yeast. The robustness of the algorithm is demonstrated in determining and maintaining the optimal dilution rate of the continuous bioreactor in presence of disturbances in environmental conditions and microbial culture characteristics. The simulation results show that a significant reduction in time required to reach optimal operating levels can be achieved using neural network model compared with the traditional dynamic linear input-output model. The extension of the algorithm for multivariable adaptive optimization of continuous bioreactor is briefly discussed.     

Finding appropriate colors for color displays

A new methodology and a supporting algorithm to determine high-contrast sets of colors for use on a color CRT are described. The method is unique in that the generated colors not only are highly contrasting, but are also ergonomically optimal. Using real data, the performance of the present method is compared to an extended version of the Carter and Carter algorithm. Finally, an integrated methodology is discussed. The latter methodology could simultaneously determine optimal phosphor luminances and chromaticities, as well as optimal values for the ambient chromaticity.