| Popis: |
To fully utilize the dynamic reconfigurability of digital microfluidic biochips, most of electrodes would be shared by different droplets. Thus, contaminations caused by liquid residues among droplets are inevitable which lead to lethal errors in bioassays. To remove the contaminations, washing operations are introduced as an essential step to ensure the correctness of bioassay. However, existing works have oversimplified assumptions on the washing droplet's behavior and constraints which cannot clean all contaminations with erroneous outcomes. Moreover, straightforward integration of washing operations with droplet routing may increase the execution time of a bioassay which is not feasible for timing-critical bioassay. To effectively remove contaminations and minimize the execution time of a bioassay, this article proposes a unified contamination-aware routing method, which addresses the above issues simultaneously. Firstly, we present a top-down scheme to generate candidates of routing paths, then construct a shortest-path model to select desirable routing solution for all subproblems. With a decision diagram of droplets, we further propose an integer linear programming (ILP) formulation to compact the execution time. Finally, contamination removal by washing droplets with realistic washing capacity is considered for all subproblems. Tested on real-life benchmarks, our proposed method can significantly reduce 72% contamination spots and save 11% execution time. |
