Improving the fine-pitch stencil printing capability using the Taguchi method and Taguchi fuzzy-based model

This paper presents industrial applications for improving the capability of the fine-pitch stencil printing process (SPP) based on the DMAIC framework and using Taguchi-based methodologies. SPP is widely recognized as the main contributor of soldering defects in a surface mount assembly (SMA). An inadequate volume of solder paste deposition or poor printing quality can cause soldering defects and lead to significant reworking and repairing costs. In practice, both the desired amount of solder paste volume (quantitative index) and printing quality (qualitative index) are preferably used to monitor the SPP for the reduction of soldering defects during the statistical control process (SPC), particularly for a fine-pitch solder paste printing operation. To continuously improve SPP capability, the DMAIC framework is followed and Taguchi-based methodologies are proposed under the considerations of single characteristic performance index (SCPI) and multiple characteristic performance indices (MCPI). The SCPI is optimized using the conventional Taguchi method. Then, a Taguchi fuzzy-based model is developed to optimize the SPP with the MCPI property. Optimizing a multi-response problem by the Taguchi method involves the engineer's judgment which tends to increase the degree of uncertainty. The performance of these two approaches is compared through the process capability metric, and the material and factors significantly affecting the fine-pitch SPP performance are reported.     

A neurofuzzy-based quality-control system for fine pitch stencil printing process in surface mount assembly

Surface mount assembly defect problems can cause significant production-time losses. About 60% of surface mount assembly defects can be attributed to the solder paste stencil printing process. This paper proposes a neurofuzzy-based quality-control system for the fine pitch stencil printing process. The neurofuzzy approach is used to model the nonlinear behavior of the stencil printing process. Eight control variables are defined for process planning and control, including stencil thickness, component pitch, aperture area, snap-off height, squeegee speed, squeegee pressure, solder paste viscosity, and solder paste particle size. The response variables are the volume and height of solder paste deposited. The values of the response variables provide indicators for identifying potential quality problems. A 3^8–3 fractional factorial experimental design is conducted to collect structured data to augment those collected from the production line for neurofuzzy learning and modeling. Visual basic programming language is then used for both rule retrieval and graphical-user-interface modeling. The effectiveness of the proposed system is illustrated through a real-world application.     

贴片机喂料器分配的优化及其遗传算法求解

针对贴片机喂料器的分配问题,给出一个新的模型,在贴装顺序已知的前提下,以贴装整块电路板所花费的总时间作为优化目标.基于该模型给出一种遗传算法,以目标函数作为其评价函数.与贪婪分配算法相比较,所花费的代价平均减少了6.2%,从而验证了该方法的有效性.     

基于强化学习的业务流程中的柔性约束研究

软件业务流程需要较高的灵活性和适应性。目前,众多的方法集中在流程建模阶段,通过一些方法提高流程的灵活性,却忽略了流程的运行阶段。由于外界环境的变化具有动态性,在建模阶段不易对其描述。由于外界环境的变化能够在运行阶段体现出来,分析流程的运行阶段是必要的。以流程中的约束为研究内容,通过对流程运行阶段的数据进行分析,提出一种基于强化学习的柔性约束模型,以提高流程的适应性。同时,将该算法应用于一类以用户为中心的复杂信息系统,实例分析表明算法是实用和有效的。     

分阶Dyna强化学习算法

针对现有Dyna强化学习体系结构下,"规划"和"学习"在计算资源分配上的不合理问题,提出了一种分阶Dyna体系结构,随着经验知识的不断积累,将学习过程划分为探索阶段、变比重学习阶段和优化阶段,分别进行"规划"和"学习"的协调控制,大大减少了计算资源的浪费.结合传统的Q-学习算法,设计了分阶Dyna-Q强化学习算法,以适应动态不确定环境下的任务.在一个标准强化学习问题中,验证了所设计的分阶Dyna强化学习算法比基本Dyna强化学习算法具有更好的学习性能.     

基于SFS原理的SMT焊点表面三维重构技术研究

基于明暗重构形状原理重构表面组装焊点的表面三维形状过程是:先通过图像采集设备,采集到SMT焊点图像,使用相关的图像处理技术,对SMT焊点图像进行处理;根据一个确定的反射模型建立物体表面形状与图像亮度之间的约束关系和物体表面形状的先验知识建立物体表面形状参数的约束关系,然后对这些约束关系联立求解,可得到物体表面的三维形状.同时针对不可接受SMT焊点图像重构出的三维图像不够理想的缺点进行了改进.在阐述其基本思想和原理的基础上,结合实例介绍了该技术的实现方法与步骤,对其中焊点图像的获取与处理、焊点三维重构技术算法等主要内容与关键技术进行了研究和探讨,并对结果进行了分析验证.     

基于神经网络的Agent增强学习模型

在深入分析Agent决策过程中状态与行为空间的迁移与构造的基础上,设计了Agent基于强化学习的最优行为选择策略和Agent强化学习的神经网络模型与算法,并对算法的收敛性进行了证明。通过对多Agent电子商务系统．中Agent竞价行为的预测仿真实验,验证了基于神经网络的Agent强化学习算法具有良好的性能和行为逼近能力。     

A reinforcement learning method based on an immune network adapted to a semi-Markov decision process

The immune system is attracting attention as a new biological information processing-type paradigm. It is a large-scale system equipped with a complicated biological defense function. It has functions of memory and learning that use interactions such as stimulus and suppression between immune cells. In this article, we propose and construct a reinforcement learning method based on an immune network adapted to a semi-Markov decision process (SMDP). We show that the proposed method is capable of dealing with a problem which is modeled as a SMDP environment through computer simulation. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

Reinforcement learning based resource allocation in business process management

Efficient resource allocation is a complex and dynamic task in business process management. Although a wide variety of mechanisms are emerging to support resource allocation in business process execution, these approaches do not consider performance optimization. This paper introduces a mechanism in which the resource allocation optimization problem is modeled as Markov decision processes and solved using reinforcement learning. The proposed mechanism observes its environment to learn appropriate policies which optimize resource allocation in business process execution. The experimental results indicate that the proposed approach outperforms well known heuristic or hand-coded strategies, and may improve the current state of business process management.     

基于小波包变换与Wiener滤波的SMT焊点图像去噪技术

针对目前SMT(surface mount technology)焊点图像去噪效果不理想的问题,提出了一种基于小波包变换与wiener滤波的SMT焊点图像去噪新方法.利用小波包对图像进行分解,可以同时对SMT焊点图像的低频和高频部分进行多层分解,有利于保留图像信息,减少噪声对图像的影响.通过对图像的小波包系数的分析,对小波包树高频系数进行Wiener滤波,保留低频系数;然后进行小波包反变换,重构得到SMT焊点去噪后图像.实验表明,提出的方法不仅可以有效地去除SMT焊点图像的噪声,而且能很好地保留原图像的边缘信息,与传统方法相比,去噪性能和去噪声效果有一定的提高.     

无模型强化学习研究综述

强化学习(Reinforcement Learning,RL)作为机器学习领域中与监督学习、无监督学习并列的第三种学习范式,通过与环境进行交互来学习,最终将累积收益最大化.常用的强化学习算法分为模型化强化学习(Model-based Reinforcement Lear-ning)和无模型强化学习(Model-free...     

融合序列模式评分的策略梯度推荐算法

推荐算法在一定程度上解决了信息过载问题,但传统推荐模型在挖掘数据特性方面有待改进。为此,结合强化学习方法提出一种融合序列模式评分的策略梯度推荐算法。将推荐过程建模为马尔可夫决策过程;分析推荐基础数据特性模式,设计以序列模式评分为奖励的反馈函数,在算法的每一次迭代过程中学习;通过对累积奖励设计标准化操作来降低策略梯度的方差。将该方法应用到电影推荐中进行验证,结果表明所提方法具有较好的推荐准确性。     

Concepts and Facilities of a Neural Reinforcement Learning Control Architecture for Technical Process Control

The paper presents the concepts of a neural control architecture that is able to learn high quality control behaviour in technical process control from scratch. As the input to the learning system, only the control target must be specified. In the first part of the article, the underlying theoretical principles of dynamic programming methods are explained, and their adaptation to the context of technical process control is described. The second part discusses the basic capabilities of the learning system on a typical benchmark problem, where a special focus lies on the quality of the acquired control law. The application to a highly nonlinear chemical reactor and to an instable multi-output system shows the ability of the proposed neural control architecture to learn even difficult control strategies from scratch.     

多步截断优先扫描强化学习算法

研究了优先扫描的强化学习方法,通过定义新的迹,把多步截断即时差分学习用于集成规划的优先扫描强化学习,用多步截断即时差分来定义扫描优先权,提出一种改进的优先扫描强化学习算法并进行仿真实验,实验结果表明,新算法的学习效率有明显的提高。     

强化学习算法中启发式回报函数的设计及其收敛性分析

(中国科学院沈阳自动化所机器人学重点实验室沈阳110016)     

强化学习中资格迹的作用

强化学习一词来自行为心理学,该学科把学习看作反复试验的过程,强化学习系统中的资格迹用来解决时间信度分配问题,文章介绍了资格迹的基本原理和实现方法。