基于蚁群算法的数字微流控生物芯片在线测试研究

The digital microfluidic biochips have a well application perspective in biochemical analysis field of gene, protein, clinical diagnosis and environmental monitoring. For the requirements of high reliability and security of chips in the application fields, so the online test is one of the important methods to ensure the normal work of the digital microfluidic biochips..Online testing is the test using the idle resources on chips in biochemical experiments, detect faults by controlling and tracking test drop routes. The test time is determined by the drop routes. This project research DMFBs online parallel test strategy, solved the effectiveness of the TSP problems based on the ant colony algorithm, applied to the test path planning of direct addressing and pin-constrained DMFBs to complete the fault detect efficiently and clear the contamination. Research the chip structure and fault model, adopts Floyd algorithm to transform the chip array into a dynamic TSP problem model with determined start and end points. Under the fluid and electrode constraints as taboo judgment strategy, the selection strategy and pheromone update mechanism of the ant colony algorithm are established to realize the test path optimization traversal and finish the fault detect. Use the ant colony algorithm to remove pollution fault, detect fluid anomaly and test fail. On the basis of clearing contamination meeting time constraints, cooperative research the fault of the chip, fluid anomaly and test fail to improve the efficiency and reliability of the test online. This test strategy can be used to both offline and online testing.

数字微流控生物芯片（DMFBs）在基因、蛋白质、临床诊断和环境监测等生化分析领域展现出良好的应用前景，由于应用领域对芯片可靠性、安全性要求高，在线测试对于确保芯片正常工作和提高工作效率异常重要。.在线测试利用芯片空闲的资源与生化实验同步进行，通过控制和追踪测试液滴的运动检测故障，液滴运动的轨迹决定测试时间的长短。本项目研究DMFBs在线并行测试策略，利用蚁群算法求解TSP问题的有效性，应用于直接寻址和引脚受限的DMFBs测试路径规划，以高效检测芯片故障、消除污染故障。通过研究芯片结构及故障模型，利用Floyd算法将芯片阵列转化为起点和终点确定的动态TSP模型，将流体和电极约束作为禁忌判断策略，建立蚁群算法的选择策略与信息素更新机制，实现测试路径的寻优遍历，完成故障检测；在增加时间约束清除污染故障的基础上，协同研究故障检测、流体异常和实验失效消除方法，提高芯片在线测试效率和可靠性。

针对数字微流控生物芯片的在线测试路径优化，课题完成了DMFBs系统测试策略的研究。针对数字微流控生物芯片的单液滴在线测试，进一步优化了灾难性故障的测试模型的转换过程，利用混合遗传蚁群算法求解测试路径，改善了单一智能优化算法收敛慢的问题，提高了算法求解测试路径的时间效率。在利用优先级策略与遗传算法进行优化时，测试路径能收敛至最优解。在此基础上，针对引脚受限的数字微流控芯片，完成了在线并行测试路径规划，减少了测试花费的时间，提高了测试效率，相比于单液滴测试，路径优化了约27%；针对引脚受限的数字微流控生物芯片的电极管脚控制信号处理进行了研究，在确保测试完成时间的情况下，减少了引脚受限DMFBs控制引脚的数量，减少了芯片的制造成本，为数字微流控生物芯片的结构设计研究提供了参考。针对数字微流控生物芯片的在线污染故障，完成了污染故障在线清除策略的研究，优化了清洗液滴的路由路径，减少了清洗时间，提高了污染故障清除效率。课题开展了数字微流控芯片在线测试研究，提高了芯片故障测试效率，确保了芯片的可靠性，为数字微流控芯片的设计与应用提供了参考。