Random permuted blocks

Abdellatif M. A SAS macro for generating randomization lists in clinical trials using permuted blocks randomization. SUGI29 Proceeding 2004 147-29. 

Block randomization (sometimes called permuted blocked randomization) addresses the issue of potential unequal distribution of subjects to the treatment groups in a simple randomization schedule by guaranteeing that the numbers of subjects allocated to each treatment are equal at the end of randomization for every block of subjects. Blocks of various sizes can be chosen, and for each block there will be a specified number of permutations for how the subjects in that block can be equally allocated to the treatment groups. The most advantageous block size for a given trial depends on the number of treatments involved (the number of treatment arms) and the numbers of subjects that participating investigative sites can recruit. 

Note that the above specification indicates that E should withstand a chosen-ciphertext attack in the preprocessing mode in the terminology of [19] or CCA-I in [3]. Possible implementation of can be done via pseudo-random permutations (which model block-ciphers). See more details on the efficient implementation of F and E in Section 4.1. 

We are presenting two major modifications of the DEW scheme here. One, energy difference is created by changes in low frequency DCT coefficients. Another is random permutation of blocks such that in any lc-region low frequency energy of lc-subregion A and that of B differ by less than some small quantity 5. Energy of a block b is the sum of absolute values of q many low frequency DCT coefficients... 

Figure 1 The "split-mix synthesis" method to generate a one-bead one-compound combinatorial library (a) and a number of permutations for random peptide libraries (b). P, , and T are building blocks (in this case amino acids).

---