Study Of Coordinated Control System Of The Biochemical Detection Process Based On Multi-Agent

Biochemical detection equipment(BDE) is a kind of clinical diagnosis instrument which is used to detect, analyze life chemistry. It can replace manual operation to complete steps of biochemical detection process like sample adding, incubation and oscillation, board washing and so on. The equipment consists of five sub devices including sample arm, mechanical arm, incubation oscillator, washing machine, detecting instrument. Through the coordinated control of the sub devices can achieve automatic operating of the detection process. Obviously, in order to improve work efficiency and automatic degree of the equipment, we must study its coordinated control process.After analyzing the functional components and operation process of BDE systematically, this paper divides the coordinated control problem of biochemical detection process into two aspects, including task scheduling optimization and dynamic coordination of equipment. Because conventional static scheduling algorithm often cannot effectively solve this kind of problem, so, this paper introduced the multi-agent technology, and put forward a coordinated control system of the biochemical detecting process based on multi-agent(MACCS). By using of the distributed solving characteristics of multi-agent, the NP hard detecting task can be decomposed into some relatively simple tasks, and then we can complete the scheduling optimization of the detection task. Through the coordination and consultation among the agents, we can achieve dynamic negotiation of the equipment, and make it working continuously and stably. Accordingly, the main research work of this paper include the following aspects:1. Analyzing the concept, characteristic, classification of agent as well as the characteristic, communication mode, cooperation mode of multi-agent system systematically, so that we can have theoretical basis to establish MACCS.2. Analyzing the key problems of BDE in depth, pointed out that the biochemical process had the characteristics of multi-constraint, dynamic and so on, then we knew the constraint conditions of the problem. We set total completion time as the optimization goal, as well as considering equipment stable and load balance, then established the MACCS model. This model uses software agent, and set the object which is involved in the biochemical detection as manage agent, coordinate agent, project agent, similar resource agent, emergency agent. Combined with the constraints and objectives, the function and structure of each agent is designed in detail.3. In order to improve its overall performance and ability of solving problems, we choose contract net protocol(CNP) as agent’s cooperation strategy. Because of the deficiencies of traditional CNP, like tender issued unrestricted, without considering the load balance and so on, we have proposed some appropriate improvement scheme. By using the earliest finish time as bidding value, the negotiation process of detection steps was discussed emphatically. Therefore, considering of the optimization problem of multi-projects sending order, a dynamic scheduling algorithm is proposed in this paper. In order to achieve dynamic negotiation of equipment, emergency agent is built specially. Through its collaboration with other agents, MACCS can handle random uncertainty like urgent task inserting, device fault, task canceling and so on.4. Making simulation examples of MACCS. Comparing multi-agent scheduling method which based on CNP and dynamic scheduling with the basic ant colony algorithm and improved ant colony algorithm and petri net scheduling method, and simulating random uncertain events including urgent task inserting, device fault, task canceling, then giving the simulating result of system’s dynamic negotiation. The simulation shows that the system can not only shorten total completion time of detection task, but also can ensure equipment’s load balance and system’s stability, and proves that MACCS has feasibility and practicability.