It is well-known that turbo equalization with the max-log-map （MLM） rather than the log-map （LM） algorithm is insensitive to signal to noise ratio （SNR） mismatch. As our first contribution, an improved MLM algor...It is well-known that turbo equalization with the max-log-map （MLM） rather than the log-map （LM） algorithm is insensitive to signal to noise ratio （SNR） mismatch. As our first contribution, an improved MLM algorithm called scaled max-log-map （SMLM） algorithm is presented. Simulation results show that the SMLM scheme can dramatically outperform the MLM without sacrificing the robustness against SNR mismatch. Unfortunately, its performance is still inferior to that of the LM algorithm with exact SNR knowledge over the class of high-loss channels. As our second contribution, a switching turbo equalization scheme, which switches between the SMLM and LM schemes, is proposed to practically close the performance gap. It is based on a novel way to estimate the SNR from the reliability values of the extrinsic information of the SMLM algorithm.展开更多
Microbial cultures are comprised of heterogeneous cells that differ according to their size and intracellular concentrations of DNA, proteins and other constituents. Because of the included level of details, multi-var...Microbial cultures are comprised of heterogeneous cells that differ according to their size and intracellular concentrations of DNA, proteins and other constituents. Because of the included level of details, multi-variable cell population balance models (PBMs) offer the most general way to describe the complicated phenomena associated with cell growth, substrate consumption and product formation. For that reason, solving and understanding of such models are essential to predict and control cell growth in the processes of biotechnological interest. Such models typically consist of a partial integro-differential equation for describing cell growth and an ordinary integro-differential equation for representing substrate consumption. However, the involved mathematical complexities make their numerical solutions challenging for the given numerical scheme. In this article, the central upwind scheme is applied to solve the single-variate and bivariate cell population balance models considering equal and unequal partitioning of cellular materials. The validity of the developed algorithms is verified through several case studies. It was found that the suggested scheme is more reliable and effective.展开更多
基金This work was supported by the National Nature Science Foundation of China under Grant No.60496313, 60502010, and 60602008.
文摘It is well-known that turbo equalization with the max-log-map （MLM） rather than the log-map （LM） algorithm is insensitive to signal to noise ratio （SNR） mismatch. As our first contribution, an improved MLM algorithm called scaled max-log-map （SMLM） algorithm is presented. Simulation results show that the SMLM scheme can dramatically outperform the MLM without sacrificing the robustness against SNR mismatch. Unfortunately, its performance is still inferior to that of the LM algorithm with exact SNR knowledge over the class of high-loss channels. As our second contribution, a switching turbo equalization scheme, which switches between the SMLM and LM schemes, is proposed to practically close the performance gap. It is based on a novel way to estimate the SNR from the reliability values of the extrinsic information of the SMLM algorithm.
文摘Microbial cultures are comprised of heterogeneous cells that differ according to their size and intracellular concentrations of DNA, proteins and other constituents. Because of the included level of details, multi-variable cell population balance models (PBMs) offer the most general way to describe the complicated phenomena associated with cell growth, substrate consumption and product formation. For that reason, solving and understanding of such models are essential to predict and control cell growth in the processes of biotechnological interest. Such models typically consist of a partial integro-differential equation for describing cell growth and an ordinary integro-differential equation for representing substrate consumption. However, the involved mathematical complexities make their numerical solutions challenging for the given numerical scheme. In this article, the central upwind scheme is applied to solve the single-variate and bivariate cell population balance models considering equal and unequal partitioning of cellular materials. The validity of the developed algorithms is verified through several case studies. It was found that the suggested scheme is more reliable and effective.
