Deterministic restrictions in circuit complexity
STOC '96 Proceedings of the twenty-eighth annual ACM symposium on Theory of computing
Group Testing Problems with Sequences in Experimental Molecular Biology
SEQUENCES '97 Proceedings of the Compression and Complexity of Sequences 1997
Improved Combinatorial Group Testing Algorithms for Real-World Problem Sizes
SIAM Journal on Computing
Explicit Non-adaptive Combinatorial Group Testing Schemes
ICALP '08 Proceedings of the 35th international colloquium on Automata, Languages and Programming, Part I
Efficiently decodable non-adaptive group testing
SODA '10 Proceedings of the twenty-first annual ACM-SIAM symposium on Discrete Algorithms
Adaptive phenotype testing for AND/OR items
ISAAC'11 Proceedings of the 22nd international conference on Algorithms and Computation
Hi-index | 0.00 |
Given n items with at most d of them having a particular property (referred as positive items), a single test on a selected subset of them is positive if the subset contains any positive item. The nonadaptive group testing problem is to design how to group the items to minimize the number of tests required to identify all positive items in which all tests are performed in parallel. This problem is well-studied and algorithms exist that match the lower bound with a small gap of log d asymptoticically. An important generalization of the problem is to consider the case that individual positive item cannot make a test positive, but a combination of them (referred as positive subsets) can do. The problem is referred as the non-adaptive complex group testing. Assume there are at most d positive subsets whose sizes are at most s, existing algorithms either require Ω(logs n) tests for general n or O((s+d dlog n) tests for some special values of n . However, the number of items in each test cannot be very small or very large in real situation. The above algorithms cannot be applied because there is no control on the number of items in each test. In this paper, we provide a novel and practical derandomized algorithm to construct the tests, which has two important properties. (1) Our algorithm requires only O ((d + s)d+s+1/(ddss) log n) tests for all positive integers n which matches the upper bound on the number of tests when all positive subsets are singletons, i.e. s = 1. (2) All tests in our algorithm can have the same number of tested items k. Thus, our algorithm can solve the problem with additional constraints on the number of tested items in each test, such as maximum or minimum number of tested items.