A Learning-Based Approach for Web Cache Management

Web caching has been widely used to alleviate Internet traffic congestion in World Wide Web (WWW) services. To reduce download throughput, an effective strategy on web cache management is needed to exploit web usage information in order to make a decision on evicting the document stored in case of cache saturation. This paper presents a so-called Learning Based Replacement algorithm (LBR), a hybrid approach towards an efficient replacement model for web caching by incorporating a machine learning technique (naive Bayes) into the LRU replacement method to improve prediction of possibility that an existing page will be revised by a succeeding request, from access history in a web log. The learned knowledge includes information on which URL objects in cache should be kept or evicted. The learning-based model is acquired to represent the hidden aspect of user request pattern for predicting the re-reference possibility. By a number of experiments, the LBR gains potential improvement of prediction on revisit probability, hit rate and byte hit rate overtraditional methods; LRU, LFU, and GDSF, respectively.     

Measurement and prediction of diffusion coefficients of supercritical CO2 in molten polymers

The solubility and diffusivity of supercritical carbon dioxide (sc‐CO₂) in low‐density polyethylene (LDPE), high‐density polyethylene (HDPE), polypropylene (PP), ethylene‐ethylacrylate copolymer (EEA) and polystyrene (PS) were measured at temperatures from 150°C to 200°C and pressures up to 12 MPa by using the Magnetic Suspension Balance (MSB), a gravimetric technique for gas sorption measurements. The solubility of CO₂ in each polymer was expressed by Henry's constant. The interaction parameter between CO₂ and polymer could be obtained from the solubility data, and it was used to estimate the Pressure‐Volume‐Temperature relationship and the specific free volume of polymer/CO₂ mixtures. The diffusion coefficients were also measured by the MSB for each polymer. The resulting diffusion coefficients were correlated with the estimated free volume of polymer/CO₂ mixture. Combining Fujita's and Maeda and Paul's diffusion models, a model was newly developed in order to predict diffusion coefficients for the polymers studied. Polym. Eng. Sci. 44:1915–1924, 2004. © 2004 Society of Plastics Engineers.     

Effect of Milk on the Deodorization of Malodorous Breath after Garlic Ingestion

Abstract: The effect of milk and milk components on the deodorization of diallyl disulfide (DADS), allyl methyl disulfide (AMDS), allyl mercaptan (AM), allyl methyl sulfide (AMS), and methyl mercaptan (MM) in the headspace of garlic as well as in the mouth- and nose-space after garlic ingestion was investigated using selected ion flow tube-mass spectrometry (SIFT-MS). Fat-free and whole milk significantly reduced the head-, mouth-, and nose-space concentrations of all volatiles. Water was the major component in milk responsible for the deodorization of volatiles. Due to its higher fat content, whole milk was more effective than fat-free milk in the deodorization of the more hydrophobic volatiles diallyl disulfide and allyl methyl disulfide. Milk was more effective than water and 10% sodium caseinate in the deodorization of allyl methyl sulfide, a persistent garlic odor, in the mouth after garlic ingestion. Addition of milk to garlic before ingestion had a higher deodorizing effect on the volatiles in the mouth than drinking milk after consuming garlic. Practical Application: Ingesting beverages or foods with high water and/or fat content such as milk may help reduce the malodorous odor in breath after garlic ingestion and mask the garlic flavor during eating. To enhance the deodorizing effect, deodorant foods should be mixed with garlic before ingestion.     

Utilization of palm oil alkyl esters as an additive in ethanol-diesel and butanol-diesel blends

This research focuses on the use of biodiesel as an additive in diesohol preparation. Three types of biodiesel—methyl, ethyl, and butyl esters—were prepared from palm oil through transesterification using a conventional base catalyst. Ethanol is generally used to blend in diesohol; however, butanol is another alcohol which has higher solubility in diesel than ethanol and it can improve the fuel properties of the blends. Therefore, a comparative study of phase stability, the dependence of solubility on temperature (10, 20, and 30 °C), and an evaluation of some basic fuel properties according to the ASTM of diesel-biodiesel-ethanol and diesel-biodiesel-butanol three-component systems at different component concentrations was done. We found that the use of butanol in diesohol could solve the problem of fuel instability at low temperatures because of its higher solubility in diesel fuel. In addition, the fuel properties results indicated that blends containing butanol have properties closer to diesel than those of blends containing ethanol.     

Preparation of carbon aerogel microspheres by a simple-injection emulsification method

Carbon aerogel microspheres were successfully prepared using a simple-injection emulsification method, employing sol–gel polycondensation of a resorcinol–formaldehyde solution containing sodium carbonate as a catalyst. This process was followed by solvent exchange using acetone, supercritical drying with carbon dioxide and carbonization in a nitrogen atmosphere. The effect of curing time before starting injection, injection rate and agitation rate of continuous phase on the particle size and the porous properties of the carbon aerogel microspheres was investigated. Adsorption of phenol by using the prepared carbon aerogel microspheres was also examined. The diameter of carbon aerogel microspheres was controlled in the range of 20–55 μm by varying injection rate and agitation rate. The mean diameter of carbon aerogel microspheres decreased with increasing the injection rate and the agitation rate, whereas their mean diameter was independent of the curing time. The BET surface area and total pore volume of carbon aerogel microspheres increased with increasing the curing time. In contrast, their BET surface area and total pore volume decreased with increasing the injection rate and the agitation rate. The BET surface area, total pore volume, mesopore volume and micropore volume of the carbon aerogel microspheres with a mean diameter of 45 μm were 903 m²/g, 0.60 cm³/g, 0.31 cm³/g and 0.27 cm³/g, respectively. The phenol-adsorption capacity of these carbon aerogel microspheres was 29.3 mg phenol/g adsorbent.     

Measurement and prediction of LDPE/CO2 solution viscosity

When CO₂ is dissolved into a polymer, the viscosity of the polymer is drastically reduced. In this paper, the melt viscosities of low‐density polyethylene (LDPE)/supercritical CO₂ solutions were measured with a capillary rheometer equipped at a foaming extruder, where CO₂ was injected into a middle of its barrel and dissolved into the molten LDPE. The viscosity measurements were performed by varying the content of CO₂ in the range of 0 to 5.0 wt% and temperature in the range of 150°C to 175°C, while monitoring the dissolved CO₂ concentration on‐line by Near Infrared spectroscopy. Pressures in the capillary tube were maintained higher than an equilibrium saturation pressure so as to prevent foaming in the tube and to realize single‐phase polymer/CO₂ solutions. By measuring the pressure drop and flow rate of polymer running through the tube, the melt viscosities were calculated. The experimental results indicated that the viscosity of LDPE/CO₂ solution was reduced to 30% of the neat polymer by dissolving CO₂ up to 5.0 wt% at temperature 150°C. A mathematical model was proposed to predict viscosity reduction owing to CO₂ dissolution. The model was developed by combining the Cross‐Carreau model with Doolittle's equation in terms of the free volume concept. With the Sanchez‐Lacombe equation of state and the solubility data measured by a magnetic suspension balance, the free volume fractions of LDPE/CO₂ solutions were calculated to accommodate the effects of temperature, pressure and CO₂ content. The developed model can successfully predict the viscosity of LDPE/CO₂ solutions from PVT data of the neat polymer and CO₂ solubility data.     

Thermoelectricity of p-Fe2O3-SO3-SiO2-others and n-Fe2O3-SiO2-CuO-others

Thermoelectric minerals have been found at Loei Province, in the northeastern part of Thailand. Local mineral specimens were prepared in the powders and bulk solids form by crushing, calcination and annealing, pressure and sintering, cutting and polishing. Mineral samples were used to analyze the composition and phase, determine the thermoelectric property and efficiency, design and construct a thermoelectric generator. Chemical composition and phase identification of powder samples were analyzed by the x-ray fluorescence (XRF) and x-ray diffraction (XRD), respectively. XRF and XRD results indicated that the mineral samples comprised the SO₃-CaO-SiO₂-others, Fe₂O₃-SO₃-SiO₂-others, Fe₂O₃-SiO₂-others and Fe₂O₃-SiO₂-CuO-others. From the thermoelectric property and efficiency determinations, the p-SO₃-CaO-SiO₂-others, p-Fe₂O₃-SO₃-SiO₂-others, n-Fe₂O₃-SiO₂-others and n-Fe₂O₃-SiO₂-CuO-others bulks were found to exhibit the thermoelectric figure of merit in orders of 10^?14, 10^?11, 10^?14 and 10^?13 K^?1, respectively. A fabricated thermoelectric generator made from ten pairs of p-Fe₂O₃-SO₃-SiO₂-others and n-Fe₂O₃-SiO₂-CuO-others legs that can be provided the open circuit voltage and short circuit current up to 48.30 mV and 0.14 μA for a temperature difference of 39.80 K at room temperature, respectively. While the internal resistance decreased and reached a value of 665 kΩ.