In this paper, a batched sparse(BATS) code transmission scheme based on chunked outer code is proposed. At the source node, a source file is first segmented into K number packets, then all the packets are divided into N chunks, each of which contains the same number of packets except for the last one if K cannot be divided by N. The chunks are encoded with the chunked outer code of BATS codes before transmitted to the next node. At the destination node, only the decodable batches need to be decoded to recover the input packets, and undecodable batches are abandoned directly to save the storage resources. In the aspect of the decoding, a simplified decoder is proposed, which is based on Gaussian elimination decoding. The consumed number of batches with a source file successfully transmitted by using the proposed scheme is analyzed under both time-invariant and time-variant channels. The simulation results show that compared with conventional BATS codes, less number of batches are consumed by applying the proposed chunked outer code.

In this paper, a batched sparse(BATS) code transmission scheme based on Low-density Parity-check(LDPC) precoding over GF(2) is studied. At the transmitter, a source file is first segmented into several packets, which are then precoded with LDPC to generate more packets. The precoded packets are encoded with the outer coding of the conventional BATS codes before they are transmitted to the intermediate nodes, which apply inner coding to the received packets and forward them to the next node. At the sink node, only a portion of received packets need to be decoded successfully to recover the source file. Moreover a one-step BP decoder is then proposed which achieves the same performance as the two-step BP decoder. The simulation results show that compared with the conventional BATS codes, less transmission batches are required by using the LDPC precoding. Meanwhile, one step decoder achieves the same transmission performance as the two step decoder, which decodes BATS codes and LDPC codes successively.

Continuously maintaining software requirements traceability links is essential for managing and evolving software systems. Due to development pressure, traceability links are usually missing during the early development phase in practice, and thus many information retrieval-based approaches have been proposed to automatically recover the traceability links. However, such approaches typically calculate textual similarities among software artifacts without considering specific features of different software artifacts, leading to less accurate results. In this paper, we propose a hybrid approach to recover requirements traceability links, which combines machine learning and logical reasoning to explore features of use cases and code. On one hand, our approach engineers features of use cases and code by taking into account their semantics, based on which a classifier is trained by using supervised learning algorithms. On the other hand, we investigate and leverage the structural information of code to incrementally discover traceability links by defining a list of reasoning rules. We have carried out a series of experiments to compare our approach with state-of-the-art methods, the results of which show that our approach significantly outperforms others.