Iterative Programming of Noisy Memory Cells

In this paper, we study a model that mimics the programming operation of memory cells. This was first introduced by Lastras-Montano et al. for continuous-alphabet channels, and later Bunte Lapidoth discrete memoryless channels (DMC). Under paradigm assume cells are programmed sequentially individually. The process is modeled as transmission over channel, such it possible to read cell state in order determine its success, case failure, reprogram again. Reprogramming can reduce bit error rate, however comes with price increasing overall time thereby affecting writing speed memory. An xmlns:xlink="http://www.w3.org/1999/xlink">iterative scheme an algorithm which specifies number attempts program each cell. Given channel constraints on average maximum cell, schemes maximize bits be reliably stored We extend results problem when either discrete-input symmetric (including BSC,BEC, BI-AWGN) or $Z$ channel. For BSC BEC our analysis also extended where probabilities consecutive writes not necessarily same. Lastly, related motivated synthesis DNA molecules.