1: \begin{abstract}

2: We consider the problem of non-adaptive noiseless group testing of $N$ items of which

3: $K$ are defective.

4: We describe four detection algorithms: the \algCOMP\ algorithm of Chan \emph{et al.}; two new algorithms,

5: \algDD\ and \algSCOMP, which require stronger evidence to declare an item defective; and an essentially

6: optimal but computationally difficult algorithm called \algSSS.

7: By considering the asymptotic \emph{rate} of these algorithms with Bernoulli designs we see that \algDD\ outperforms

8: \algCOMP, that \algDD\ is essentially optimal in regimes where $K \geq \sqrt N$, and that no algorithm

9: with a nonadaptive Bernoulli design can perform as well as the best non-random adaptive designs when

10: $K > N^{0.35}$. In simulations, we see that \algDD\ and \algSCOMP\ far outperform \algCOMP, with

11: \algSCOMP\ very close to the optimal \algSSS, especially in cases with larger $K$.

12: %

13: %

14: %We develop several new algorithms

15: %and analyse their probability of success. In particular, we describe the Definite Defectives

16: %(\algdd\!\!) algorithm, and show analytically that is asymptotically

17: %close to optimal in regimes where the number of defective items $K \geq \sqrt N$, where $N$ is the total number of items. Further, we describe the \algscomp algorithm, and 

18: %perform

19: %simulations which show that it performs even better than the \algdd algorithm. Finally, we derive upper bounds on the

20: %success probability for any algorithms based on a Bernoulli sampled test matrix. 

21: %We analyse the normalized number of tests required, in terms of information theoretic capacity.

22: %In certain regimes we argue that in an asymptotic sense the \algdd algorithm is optimal, and that

23: %less sparse problems (more defectives) have a larger `adaptivity gap'.

24: \end{abstract}

25: