基于启发式Q(λ)学习的铁路绝缘子定位研究

智能化冲洗是解决电气化铁路绝缘子人工冲洗弊端的有效途径,而智能化冲洗的关键在于绝缘子识别、定位。通过对现有智能冲洗设备的分析,提出一种铁路绝缘子定位、跟踪方法,该方法基于传感器信息,采用启发式Q(λ)学习算法,可快速、准确找到绝缘子。首先建立铁路绝缘子水冲洗环境模型,接着为解决传统Q(λ)学习算法盲目学习导致收敛速度慢的问题,通过搜索支柱特征调整奖赏函数,设计启发式策略函数并融入Q(λ)学习中,提高算法的学习和收敛速度。最后分别对传统Q(λ)学习算法和启发式Q(λ)学习算法进行Matlab仿真实验,确定最佳参数设置,仿真结果表明启发式Q(λ)算法的正确性和可行性。     

Human myeloma light chains with increased molecular weight: high frequency among lambda chains

The discovery of a human myeloma protein comprising a kappa L-chain with an increased mol. wt of 30,000) (Bouvet et. al., 1980) prompted investigations on the incidence of such heavier L-chains among other human myeloma proteins. In 105 samples examined, 34 were found to have L-chains heavier than normal (23,000-24,000), ranging from 25,000 up to 31,000, and five of lighter mol. wt (21,000-22,000). These mol. wt abnormalities were detected by electrophoresis in sodium dodecyl sulfate 10% polyacrylamide gels (SDS-PAGE) after reduction with 2-mercaptoethanol. The mol. wt of three of the heavier kappa or lambda chains was also estimated by filtration through a Sephadex G100 column and by sedimentation equilibrium. All three methods indicated a mol. wt increase of about 15-25% as compared with the usual mol. wt. The distribution of the high mol. wt chains among all L-chains examined was found to be 11 out of 62 kappa chains (17.7%) and 23 out of 43 lambda chains (53%) (P less than 0.001). A preferential association of such L-chains with H-chains producing multiple bands in SDS-PAGE (P less than 0.01) and an association between multiple L-chain and multiple H-chain band (P less than 0.05) were also observed. In contrast, no abnormal L-chain was found in immunoglobulins from normal subjects. Spontaneous degradation of the normal H-chains sometimes yielded fragments of 30,000 mol. wt. These fragments were easily distinguishable from abnormal L-chains. The nature of extra mol. wt in heavy L-chains was investigated for the presence of carbohydrate moiety. Four large and three normal size L-chains were examined for amino-sugar and sialic acid content. A small amount (one residue per molecule) of amino-sugar was detected only in two normal and two heavy L-chains, whereas sialic acid was only found in the heaviest (27,000-30,000) L-chains (Lh) and in small percentage (one or two residues per molecule). Total sugar estimation in one Lh chain indicated a proportion not exceeding three or four residues per L-chain (mol. wt 1,000) and this is insufficient to explain the 15-25% (3,600-6,000) mol. wt increase. It is therefore possible that, at least in some heavy myeloma L-chains, an additional peptide is expressed. Whatever the nature of the increase it would be of interest to elucidate whether this is a marker of malignant process or of an intermediate step of normal Ig synthesis.     

精馏塔时滞过程的分数阶PIλDμ控制器设计

本文针对精馏塔不稳定时滞过程,提出一种基于设定值加权的分数阶PIλDμ控制器设计方法.首先采用比例环节构成内环反馈镇定不稳定时滞过程,然后基于等效的过程模型,依据设定值加权方法设计分数阶PIλDμ控制器,并进行了控制器参数整定.仿真结果表明:分数阶PIλDμ控制器可以使精馏塔不稳定时滞过程系统获得良好的动态响应特性,干扰抑制特性以及克服系统参数变化的鲁棒性.     

基于蕴涵算子族L-λ-R_0的模糊推理α-三I约束算法

提出了基于蕴涵算子族L-λ-R0的模糊推理的思想,这将有助于提高推理结果的可靠性。针对蕴涵算子族L-λ-R0给出了模糊推理的FMP模型及FMT模型的α-三I约束算法。     

基于蕴涵算子族Lλ0λG的反向三I支持算法

提出了基于蕴涵算子族Lλ0λG的模糊推理的思想,这将有助于提高推理结果的可靠性。针对蕴涵算子族Lλ0λG给出了模糊推理的FMP模型及FMT模型的反向三I支持算法。     

利用标记基因选配褐壳蛋鸡配套杂交亲本

应用本实验室研制的抗鸡红细胞抗原单价血清(4个位点, 14个等位基因)和DNA指纹技术,对我们组配成功的一个褐壳蛋鸡配套系统的5个亲本进行了群体遗传学分析。结果表明,由标记基因测定所提供的亲本品系遗传差异的大小, 与这些品系实际杂交效果的优劣相一致,证实了标记辅助选种方法有的效性。     

Robust skeletonization using the discrete λ-medial axis

Medial axes and skeletons are notoriously sensitive to contour irregularities. This lack of stability is a serious problem for applications in e.g. shape analysis and recognition. In 2005, Chazal and Lieutier introduced the λ-medial axis as a new concept for computing the medial axis of a shape subject to single parameter filtering. The λ-medial axis is stable under small shape perturbations, as proved by these authors. In this article, a discrete λ-medial axis (DLMA) is introduced and compared with the recently introduced integer medial axis (GIMA). We show that DLMA provides measurably better results than GIMA, with regard to stability and sensibility to rotations. We give efficient algorithms to compute the DLMA, and we also introduce a variant of the DLMA which may be computed in linear-time.     

Molecular cloning of bovine leukemia virus DNA integrated into the bovine tumor cell genome

The bovine leukemia virus (BLV) DNA harbored in the bovine tumor cell genome was cloned in lambda Charon 4A phage. Using either representative or 3' half-enriched BLV cDNA as a blot hybridization probe, clone lambda BLV-1 was shown to carry 9 kb of the BLV genome, flanked by cellular sequences at both ends. Restriction mapping with twelve endonucleases and hybridization of the DNA fragments to BLV cDNA representing a 3'-end portion of the viral genome revealed the presence and precise location of two long terminal repeats (LTRs) and virus-cell junctions. Thus, lambda BLV-1 appears to contain the complete BLV genome and flanking tumor cellular sequences. The restriction map of the cloned BLV proviral DNA closely resembles that previously reported for unintegrated linear proviral DNA, but differs significantly from that of the integrated provirus of another BLV isolate, the difference occurring preferentially in the putative gag and pol genes.     

一种非线性模型参考迭代学习控制算法

针对存在周期或非周期不确定干扰的非线性系统,提出一种模型参考迭代学习控制算法(MR-ILC),其目标是在被控系统满足假设条件并采用相应学习律的条件下,使被控系统的状态变量跟踪参考模型的状态变量,且两种模型不需要具有相同的结构和参数。运用λ范数证明了算法的收敛性,仿真结果验证了算法的有效性。