• Journal of the Korea Society of Computer and Information
• /
• v.16 no.2
• /
• pp.93-99
• /
• 2011

In this paper, we propose a policy-gradient routing scheme under Reinforcement Learning that can be used adaptive QoS routing. A policy-gradient RL routing can provide fast learning of network environments as using optimal policy adapted average estimate rewards gradient values. This technique shows that fast of learning network environments results in high success rate of routing. For prove it, we simulate and compare with three different schemes.

• Journal of the Korean housing association
• /
• v.1 no.1
• /
• pp.9-26
• /
• 1990

The purpose of this study is to analyze the quantity, residents` storing manner, and frequency in use of the stored items in bedrooms. and to comprehend the relevant desires of the residents in apartments. The results of the study are summarized as follows; 1. The quantity of the stored Items in bedrooms of apartments were measured after classified into 4 categories; Western style clothes, Korean costumes, underwears, and miscellaneous items. The quantity of the stored items In Rl(the largest bedrooms) was more than twice as many as in R2(the second largest bedrooms)2. The stored items were divided inte 3 categories according to the manner in which they were stored; those generally hung, those generally folded, and those hung or folded half-and-half.3. The frequency in use of the stored items was classified into 2 categories-those used "always" or "often" and those used "seasonal" or "seldom"·and shown diversely according to the kind of items4. It was shown that most residents in apartments were not much satisfied with storage space, i.e.. they generally fell the storage space insufficient. And they tended to want more wardrobes in Rl, and more shelves in R2.

• The Journal of Information Technology
• /
• v.5 no.1
• /
• pp.1-12
• /
• 2002

This paper presents a composition method for document image using RDM algorithm. It is possible to compose about double quantity of document image in same document space compared with RL or DM algorithm, if it used. RDM algorithm is available to compose secret document as well as digital signatures onto non-secure document. In this case, secure transmissin of document will be realize because the third party do not recognize secure transmission.

• Journal of Food Hygiene and Safety
• /
• v.14 no.2
• /
• pp.134-139
• /
• 1999

The relationship between storage stability and chlorophyll contents of dried laver(DL), roasted laver(RL) and roasted seasoned laver(RSL) according to packaging during storage at 20$\pm$0.5$^{\circ}C$ were investigated by measuring the acid value(AV) and the peroxide value(POV). There results are as follows; 1.The storage stability was decreased in the rank of 0 < 3< 6 < 9 months by storage term, RL < DL < RSL by the kind of laver products and PP/PE/AL/PE/LLDPE(Al) < PP/PE(PP) y packaging. 2. When the silica gel was added to RSL, the storage stability was improved. The more the addition level of silica gel was increased, the more the storage stability was improved. The rank order was 2g < 4g in RSL. 3. Reduction of the total chlorophyll content in RSL were deacreased to 10% when packed with Al and 24% when packed with PP.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.52 no.9
• /
• pp.533-539
• /
• 2003

Reinforcement learning (RL) has been widely used as a learning mechanism of an artificial life system. However, RL usually suffers from slow convergence to the optimum state-action sequence or a sequence of stimulus-response (SR) behaviors, and may not correctly work in non-Markov processes. In this paper, first, to cope with slow-convergence problem, if some state-action pairs are considered as disturbance for optimum sequence, then they no to be eliminated in long-term memory (LTM), where such disturbances are found by a shortest path-finding algorithm. This process is shown to let the system get an enhanced learning speed. Second, to partly solve a non-Markov problem, if a stimulus is frequently met in a searching-process, then the stimulus will be classified as a sequential percept for a non-Markov hidden state. And thus, a correct behavior for a non-Markov hidden state can be learned as in a Markov environment. To show the validity of our proposed learning technologies, several simulation result j will be illustrated.

• Archives of Pharmacal Research
• /
• v.13 no.4
• /
• pp.293-297
• /
• 1990

Sustained release microspheres containing phenylpropanolamine HCI (PPA) were prepared with acrylic polymer (Eudragit RL/RS) sand hydroxypropylmethylcellulose phthalate (HPMCP) using a emulsion-solvent evaporation method. Magnesium strate was used a smoothing agent for preparation of microspheres. The microspheres obtained were very spherical and free-flowing particles. Scanning electron microscopy showed that microspheres have a smooth surface and a sponage-like internal structure. The dissolution rate of PPA from the microspheres was dependent on the pH of dissolution media. PPA showed faster relase in hP 1. 2 solution than in pH 7.4 solution due to the solubility of PPA. Therefore we prepared new microspheres containing 5% (w/v) HPMCP in order to control the release of PPA. The release rate of PPA from these new microspheres was similar in pH 1.2 and pH 7.4 solution.

• Archives of Pharmacal Research
• /
• v.29 no.10
• /
• pp.921-927
• /
• 2006

Prolonged release micro particles of clarithromycin (CL) were prepared using Eudragit RL 100 and RS 100 by spray-drying and casting-drying techniques. For the characterization of those microparticles, preparation yield, particle size distribution, X-ray diffraction, thermal behavior, active agent content and in vitro dissolution from the microparticles were performed. HPLC was used for the assay of clarithromycin and the assay method was validated. All the formulations obtained showed prolonged release when compared to pure clarithromycin. Microparticles prepared by spray-drying method had a slower release compared to those of casting drying method. Spray-drying method seems to be a more suitable method to prepare microparticles for prolongation in release.

• Journal of Power Electronics
• /
• v.16 no.3
• /
• pp.1122-1130
• /
• 2016

This paper presents a novel approach for achieving both a tight DC voltage regulation and a power factor control by applying the Reaching Law Sliding Mode Control (RL-SMC) and the conventional Sliding Mode Control (SMC). Applying these strategies on a matrix rectifier (MR) can achieve a unity grid side power factor when the DC load changes widely and it can provide a ripple-free output voltage that is easily affected by distortions of the three-phase ac voltage supply. Furthermore, by employing the reaching law on the SMC can solve the chatting problem of the sliding motion. Comparative Matlab simulations and experimental verifications for these strategies have been presented and discussed in this paper. The results show that by applying the SMC and RL-SMC on a MR can achieve a unity grid side power factor and a regulated ripple-free DC output.

• KSBB Journal
• /
• v.13 no.2
• /
• pp.133-140
• /
• 1998

Using rlFN-$\alpha$ and rhGH as the model proteins, the refolding performances of the published processes were evaluated and compared. Key engineering parameters such as the type of denaturant and this concentration, protein concentration in the refolding buffer, and pH and ionic strength of the buffer were experimentally investigated. Furthermore, the role of a co-solvent of surfactant type in aggregation reduction was also studied. Of the denaturants tested (8M urea, 6M guanidine HCI, 0.5% SDS), SDS at alkaline pH (9.5) and ambient temperature gave the highest recovery yield. The SDS process was effective in the refolding of observed where dissolution proceeded better under lower strength (10 mM) but aggregation was suppressed under higher strength (>50 mM.) When PEG-4000 and/or Tween were added as co-solvent or refolding-enhancing additive, 1.6-2 times higher yield was realized. The‘masking’of the hyrophobic patches located on the surface of the protein with the surfactant molecules was believed to be responsible for the considerable reduction in aggregation during refolding.

• Journal of Ginseng Research
• /
• v.19 no.3
• /
• pp.231-236
• /
• 1995

The antioxidant activity and the pharmacologically active constituents such as nucleosides and phenolic substancs of Panax ginseng C.A. Meyer (white ginseng powder) treated with gamma irradiation (Rl group) and ethylene oxide fumigation (EO group) were compared with those of non-treated ginseng (control group). High performance liquid chromatograms of nucleosides in three groups were similar with each other. The bathochromic effect of phenolic substances at W absorption region tends to be higher for EO group than Rl group. This tendency was also observed in heat treatment of each group. However, the antioxidant activity may have some effects on the labile property of phenolic substances. In summary the gamma-irradiated ginseng was more stable than the ethylene oxide-treated ginseng, but both ginseng samples gave the similar antioxidant activity.