全自动生化分析仪多任务优化调度与检测方法研究

发布时间：2018-04-08 19:07

本文选题：全自动生化分析仪　切入点：遗传算法　出处：《湖南大学》2012年硕士论文

【摘要】：生化分析仪作为一种主要医疗仪器通过采集的血液或体液样本完成人体中各种生化指标（如血糖、血脂、胆红素及各种催化酶）的定量分析，检测肝功、肾功、心肌酶谱、电解质等健康状况。目前，国内绝大多数医院主要采用国外进口生化分析仪，具有检测精度高、稳定性强、自动化程度高、功能强大等诸多优点。随着技术的进步，国外一些生化分析仪制造厂家研发出了大型全自动生化分析仪，能够完成多指标及复杂任务的快速检测，无需人工参与。相比国外发展状况，国内全自动生化分析仪的发展较为缓慢，虽然已实现生化分析基本功能，但却有诸多不足之处，主要体现在以下几个方面：（1）国内全自动生化分析仪所需光源、高精密吸液机构等核心元器件主要依靠进口，尚无自主研发能力；（2）生化分析检测方法简单，生化分析结果精度低、重复性和稳定性差，常需人工参与，进行重复或稀释测试；（3）生化分析调度自动化程度低，常采用人工设定的固定周期进行项目调度，导致单次生化分析时间长、多任务并行检测技术瓶颈难以突破。本文根据目前国内大中型生化分析仪研发过程中存在的上述问题，进行了系统深入研究，取得的工作和创新具体如下：针对用动力学法测定酶活力过程中遇到的底物耗尽问题，建立新的数据处理模型，运用所提出的新算法自动判别数据的线性区间与非线性区间，自动选择合适的线性数据用于酶活力的计算，，从而提高仪器的工作效率，降低重复测试的次数，扩大仪器对样品测试的浓度范围，进一步加快了使用的工作效率。针对目前全自动生化分析仪中对调度问题采用的最大固定周期流水作业方法中存在检测时间长、效率低、存在间歇性和不连续性等诸多问题，和已有的基于Job-Shop生化调度方法中易搜索至局部最优和搜索效率不高的问题，提出了一种基于遗传算法寻优的非固定周期多任务调度方法。该方法打破了按机械位置次序依次检测的执行方式，以批次处理时间最短为原则，设计基于grefenstette编码的遗传算法ATSP模型，优化交叉和变异算子，解决在交叉和变异过程中产生非法路径的难题，建立全自动生化分析仪的ATSP模型，优化全自动生化分析仪调度。本文在Delphi环境下实现了所提算法和新的数据处理模型的建立，采用SQLSever完成了数据库模块的设计，对生化分析仪的软、硬件系统进行了详细的介绍，并基于该平台完成对本文所提方法的验证，较好的完成预期目标。
[Abstract]:Biochemical analyzer as a major medical instrument by collecting blood or body fluid samples to complete various biochemical indexes of body (such as blood glucose, blood lipids, bilirubin and various enzymes) quantitative analysis and detection of liver function, renal function, myocardial enzymes, electrolytes and other health conditions. At present, most domestic hospitals mainly using imported biochemical analyzer has high detection precision, strong stability, high degree of automation, powerful advantages. With the development of technology, some foreign manufacturers of biochemical analyzer developed large-scale automatic biochemical analyzer, rapid detection of multi index and able to complete complex tasks, without artificial participation. Compared the development situation of foreign and domestic development automatic biochemical analyzer is relatively slow, although the realization of the basic functions of biochemical analysis, but there are many shortcomings, mainly reflected in the following aspects: (1) China In the automatic biochemical analyzer for light source, the core components of high precision suction mechanism mainly rely on imports, there is no independent research and development ability; (2) the biochemical analysis of biochemical analysis of the detection method is simple, low precision, repeatability and stability is poor, often need artificial participation, repeat or dilution test; (3) the biochemical analysis of dispatching automation low level, often using fixed cycle artificial set of project scheduling, resulting in a single biochemical analysis of long time, multi task parallel bottleneck detection technology is difficult to break.
Based on the above problems existing in the development process of large and medium-sized biochemical analyzers in China, this paper makes a systematic and in-depth research.
According to the kinetic method for the determination of enzyme activity in the process of encounter substrate exhaustion problem, build a new data processing model, a new algorithm using the proposed automatic identification of linear and nonlinear interval interval data calculation, automatic selection of linear data suitable for enzyme activity, so as to improve the work efficiency of the instrument, reduce the number of repeated testing, expand the concentration range of the instrument on the sample test, and further accelerate the use efficiency.
Aiming at the detection time of maximum fixed cycle flow current methods of automatic biochemical analyzer on the scheduling problem in the long, low efficiency, intermittency and discontinuity problems based on Job-Shop biochemical scheduling method and the easy to search the local optimum and the search efficiency is not high, this paper proposes a fixed cycle multi task scheduling method based on genetic algorithm optimization. The method breaks the order according to the mechanical position detection means of implementation, to batch the principle of shortest time, genetic algorithm design of ATSP model based on grefenstette encoding optimization, crossover and mutation operators, solve the problem of generating illegal path in the process of crossover and mutation the establishment of ATSP model, automatic biochemical analyzer, automatic biochemical analyzer to optimize the scheduling.
In the Delphi environment to achieve the proposed algorithm and a new data processing model, using SQLSever to complete the design of the database module, the biochemical analyzer software and hardware system are introduced in detail, and based on the platform to complete the verification of the proposed method, accomplish the expected goal.

【学位授予单位】：湖南大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R318.6

【参考文献】