• Nuclear Engineering and Technology
• /
• v.54 no.10
• /
• pp.3864-3877
• /
• 2022

We propose a new approach called PESA (Prioritized replay Evolutionary and Swarm Algorithms) combining prioritized replay of reinforcement learning with hybrid evolutionary algorithms. PESA hybridizes different evolutionary and swarm algorithms such as particle swarm optimization, evolution strategies, simulated annealing, and differential evolution, with a modular approach to account for other algorithms. PESA hybridizes three algorithms by storing their solutions in a shared replay memory, then applying prioritized replay to redistribute data between the integral algorithms in frequent form based on their fitness and priority values, which significantly enhances sample diversity and algorithm exploration. Additionally, greedy replay is used implicitly to improve PESA exploitation close to the end of evolution. PESA features in balancing exploration and exploitation during search and the parallel computing result in an agnostic excellent performance over a wide range of experiments and problems presented in this work. PESA also shows very good scalability with number of processors in solving an expensive problem of optimizing nuclear fuel in nuclear power plants. PESA's competitive performance and modularity over all experiments allow it to join the family of evolutionary algorithms as a new hybrid algorithm; unleashing the power of parallel computing for expensive optimization.

• Korean Journal of Ecology and Environment
• /
• v.46 no.2
• /
• pp.215-224
• /
• 2013

We clarified the background for establishment of vegetation by comparing the spatial distribution maps of vegetation and substrate on a gravel bar in the Bangtae stream located on Inje-gun of Gangwon-do, the central eastern Korea. The total vegetation coverage was higher in the interior and lower in the marginal parts of the gravel bar. Spatial distribution of vegetation on the longitudinal section of the gravel bar tended to be arranged in the order of shrub, subtree, and tree dominated vegetation types from the front (upstream) toward the rear (downstream) parts. Coverage of the herbaceous plants was higher in the central and rear parts and lower in the front and right parts of the gravel bar. Vegetation height was higher in the rear part and became lowered as move toward the front part. Substrate was distributed in the order of boulder, gravel, sand, and boulder from the front toward the rear parts. Ordination of stands based on vegetation data was arranged in the order of annual plant, perennial herb, shrub, and tree dominated vegetation as move from the right to the left parts on the axis I. Species richness was higher in the order of Pinus densiflora community, Phragmites japonica community, Salix gracilistyla community, Fraxinus rhynchophylla community, annual plant dominated vegetation, and Prunus padus for. padus community based on the species rank-abundance curve. The order based on the Shannon's index was some different; diversity of Phragmites japonica community and Salix gracilistyla community, which showed higher dominance degree, were low differently from species richness. In conclusion, it was evaluated that the gravel bar newly established toward the upstream and vegetation dynamics of the gravel bar seemed to follow ecosystem mechanisms of succession. As were shown in the above results, the Bangtae stream corresponded to the upstream and thereby particle size of substrate was big. Therefore, they move by rolling and are accumulated for the upstream. Vegetation types were arranged in the order of woodland, shrub-land and grassland from the rear toward the front parts of the gravel bar and thereby reflected the formation process of the bar. However, the gravel bar is disturbed frequently by not only the running water but also the suspended sand as the dynamic space. Such disturbances cause habitat diversity and consequently led to high biodiversity.

• Journal of Life Science
• /
• v.13 no.4
• /
• pp.541-550
• /
• 2003

E2 glycoprotein of hepatitis C virus (HCV) comprises a surface of viral particle together with E1 glycoprotein, and is thought to be involved in the attachment of HCV viral particle to receptor (s) on the permissible cells including hepatocytes, B cells, T cells, and monocytes. We constructed a phage library expressing cellular proteins of hepatocytes on the phage surface, which turned out to be 8.8${\times}$$10^5$ cfu of diversity and carried inserts in 95% of library. We screened both cDNA phage library and 12-mer peptide library to identify the cellular proteins binding to E2 protein. Some intracellular proteins including tensin and membrane band 4.1 which are involved in signal transduction of survival and cytoskeleton organization, were selected from cDNA phage library through several rounds of panning and screening. On the contrary, membrane proteins such as CCR7, CKR-L2, and insulin-like growth factor-1 receptor were identified through screening of peptide library. Phages expressing peptides corresponding to those membrane proteins were bound to E2 protein specifically as determined by neutralization of binding assay. Since it is well known that HCV can infect T cells as well as hepatocytes, we examined to see if E2 protein can bind to CCR7, a member of C-protein coupled receptor family expressed on T cells, using CCR7 transfected tells. Human CCR7 cDNA was cloned into pcDNA3.1(-) vector and transfected into human embryonic kidney cell, 293T, and expressed on the surface of the cell as shown by flow cytometer. Binding assay of E2 protein using CCR7 transfected cells indicated that E2 protein bound to CCR7 by dose-dependent mode, giving rise to the possibility that CCR7 might be a putative cellular receptor for HCV.