移動規劃之複雜度問題

The coronavirus COVID-19 is affecting most countries and territories around the world. Efficient clinic tests are fundamental to the ability to detect and respond. Pooling methods are therefore proposed to improve the performance of diagnostic tests. This paper examines how to retrieve those samples that are positive for COVID-19 while performing only a small number of tests: the samples are arranged on a 2-dimensional grid and pooled according to the rows, columns, and cyclic diagonals. According to the outcomes, a fingerprint is determined for each sample, allowing the selection of a list containing most of the positive samples with a few false positives and a list of uncertain samples. These uncertain samples are further filtered in the second stage by performing individual-testing. The advantage of the proposed method is the low number of pools without reducing the high quality of the diagnosis. Intuitively, the drawback of the pooling strategy compared to individual-testing (non-pooling) approaches is that it raises the possibility of false negatives. Our method, however, leads to a comparable result with a low rate of error, even better than that of the individual-testing strategy in some cases. Furthermore, this method allows an easy and systematic screening and can be applied to a mass testing project. Such an implementation will be a straightforward compromise between sensitivity and efficiency to expand current screening capacities, thereby be well-suited to mass testing for potential asymptomatic and mild cases.