Research And Implementation Of Automatic Biochemical Analyzer Parallel Scheduling Algorithm

With the continuous progress of modern science and technology and the rapid development of medical level, biochemical analyzer has been developed into a comprehensive system.which for obtaining specific substance concentration information. It is available to the body fluid samples for biochemical and electrolyte detection project. At present, the highest detection rate of biochemical analyzer test with 2000 / hour. In order to ensure the precision and accuracy of sampling, the hardware has been difficult to further improve the detection speed. The software task scheduling algorithm is one of the main factors affecting the biochemical detection rate. How to use more efficient scheduling algorithm to improve the detection efficiency of biochemical analysis system has become a major problem to limit its development.The main work of this paper can be summarized as the following aspects:(1) By adding the special restriction conditions, we have transformed the task scheduling problem of the biochemical analyzer. Biochemical analyzer task scheduling problem can be understood as pipeline operations under a special case, the problems in the widely used fixed cycle scheduling algorithms have low detection efficiency, waiting for long time, by adding a virtual task, establishing the task start time matrix, biochemical analyzer task scheduling special pipeline operations that casts the problem for a non symmetric type traveling salesman problem, and the establishment of related task initiation form.(2) In view of the transformation of mathematical problems, the neural network algorithm model is constructed to solve the problem. The optimal task scheduling order is solved by using the neuron state and the energy matrix of the neural network. Combined with greedy algorithm is obtained by the iterative algorithm of neural network of neurons in the total energy matrix local optimization, two-dimensionalmatrix problems fast conversion for one-dimensional problems, to solve the single neural network algorithm in the presence of backtracking for long periods of time. According to the specific test task, the neural network- greedy algorithm is established in the mathematical model of the scheduling algorithm on the full automatic biochemical analyzer.(3) In this paper, the scheduling algorithm and the commonly used fixed cycle scheduling algorithm were simulated in the MATLAB platform for the same task scheduling problem, we obtained two different algorithms under the time-consuming task Gantt chart. Eight sets of simulation tests were carried out, and the average velocity was 31%.(4) On the Microsoft Visual Studio 2013 platform, we realized the application of the proposed scheduling algorithm and the fixed cycle scheduling algorithm of the biochemical analysis software control system. A sample test experiment is carried out under the two algorithms. After eight groups of experiments, the rate of increase was 24%. Fixed cycle algorithm to detect the speed of 400 test / hour, the neural network- greedy algorithm to detect the speed of 496 test / hour.In this paper, we present an efficient task scheduling algorithm and set up its mathematical model in the biochemical analysis system. Compared with the traditional fixed cycle scheduling algorithm, the simulation results are compared and analyzed. On this basis, we realize the application of this scheduling algorithm and the fixed cycle algorithm of biochemical analysis software system platform. The superiority of the scheduling algorithm is verified by experiments.