Research On Control-Fluidic Physical CoDesign Algorithm For Microfluidic Biochips

With the quick development of microfluidic biochips,a trend of increasing integration density has been observed on microfluidic biochips.The manual design can not meet the requirements of quality and efficiency any more.Furthermore,as the bioassays scenarios and user requirements become complex and diverse,the number of components and reaction agents are increasing dramatically,which leads to a series of controlling and scheduling issues.All these issues cause great challenges on design automation of microfluidic biochips.Meanwhile,there are two major design objectives:(1)Decreasing the fabrication cost of biochips,and(2)Increasing the reliability of biochips.This thesis focuses on three different types of microfluidic biochips and proposes a series of effective and efficient algorithms.Main contributions of this thesis are as follows:Firstly,a droplet and electrode oriented physical codesign method is proposed to solve the design gap issues of flow-layer design and control-layer design in digital microfluidic biochips.In flow-layer design stage,based on A searching and topological sorting algorithms,the contamination issue with washing capacity constraint is solved.In control-layer design stage,based on support vector machine,the design gap issues between electrode clustering and control line routing are solved by feature extraction and training of classification model.Experimental results show that the proposed method has better performance on electrode connection success rate and the number of used electrodes.And the proposed method obtains more than 30%speedups compared with the existing method.Furthermore,the washing capacity constraint is considered for the first time in the proposed method.Secondly,the physical codesign method for continuous-flow microfluidic biochips is proposed.Based on sequence-pair representation,negotiation-based strategy,network-flow formulation,and iterative-feedback mechanism,the design gap issues between flow-layer design stage and control-layer design stage are solved.Furthermore,a pressure-aware control layer optimization method is proposed to solve the pressure degradation issue when multiplexer is adopted.Experimental results show that the proposed method imporves the chip area,the number of flow channel intersections,and the total flow channel length on average by 32%,40%,and 22%,respectively.Furthermore,the proposed method can solve the pressure degradation issue and obtains an average 77%improvement in total refreshing cost compared with the baseline methodThirdly,bioassays oriented physical codesign method for paper-based digital mi-crofluidic biochips is proposed.The control interference issue is first considered in this work.The network-flow formulation and integer linear programming are adopted in the proposed control-fluidic physical codesign method.The control interference issue is solved by iterative-feedback mechanism.Experimental results show that the proposed method obtains an average 41%improvement in electrode connection success rate com-pared with the baseline method.Furthermore,the electrode-oriented and droplet-oriented codesign methods are specially designed for significantly enhanced efficiency for different types of bioassays.