下载文档原格式
HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationFLUOXETINE HClC17H18F3NO•HClM.W. = 345.79CAS — 59333-67-4STABILITY INDICATINGA S S A Y V A L I D A T I O NMethod is suitable for:ýIn-process controlþProduct ReleaseþStability indicating analysis (Suitability - US/EU Product) CAUTIONFLUOXETINE HYDROCHLORIDE IS A HAZARDOUS CHEMICAL AND SHOULD BE HANDLED ONLY UNDER CONDITIONS SUITABLE FOR HAZARDOUS WORK.IT IS HIGHLY PRESSURE SENSITIVE AND ADEQUATE PRECAUTIONS SHOULD BE TAKEN TO AVOID ANY MECHANICAL FORCE (SUCH AS GRINDING, CRUSHING, ETC.) ON THE POWDER.ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationTABLE OF CONTENTS INTRODUCTION........................................................................................................................ PRECISION............................................................................................................................... System Repeatability ................................................................................................................ Method Repeatability................................................................................................................. Intermediate Precision .............................................................................................................. LINEARITY................................................................................................................................ RANGE...................................................................................................................................... ACCURACY............................................................................................................................... Accuracy of Standard Injections................................................................................................ Accuracy of the Drug Product.................................................................................................... VALIDATION OF FLUOXETINE HCl AT LOW CONCENTRATION........................................... Linearity at Low Concentrations................................................................................................. Accuracy of Fluoxetine HCl at Low Concentration..................................................................... System Repeatability................................................................................................................. Quantitation Limit....................................................................................................................... Detection Limit........................................................................................................................... VALIDATION FOR META-FLUOXETINE HCl (POSSIBLE IMPURITIES).................................. Meta-Fluoxetine HCl linearity at 0.05% - 1.0%........................................................................... Detection Limit for Fluoxetine HCl.............................................................................................. Quantitation Limit for Meta Fluoxetine HCl................................................................................ Accuracy for Meta-Fluoxetine HCl ............................................................................................ Method Repeatability for Meta-Fluoxetine HCl........................................................................... Intermediate Precision for Meta-Fluoxetine HCl......................................................................... SPECIFICITY - STABILITY INDICATING EVALUATION OF THE METHOD............................. FORCED DEGRADATION OF FINISHED PRODUCT AND STANDARD..................................1. Unstressed analysis...............................................................................................................2. Acid Hydrolysis stressed analysis..........................................................................................3. Base hydrolysis stressed analysis.........................................................................................4. Oxidation stressed analysis...................................................................................................5. Sunlight stressed analysis.....................................................................................................6. Heat of solution stressed analysis.........................................................................................7. Heat of powder stressed analysis.......................................................................................... System Suitability stressed analysis.......................................................................................... Placebo...................................................................................................................................... STABILITY OF STANDARD AND SAMPLE SOLUTIONS......................................................... Standard Solution...................................................................................................................... Sample Solutions....................................................................................................................... ROBUSTNESS.......................................................................................................................... Extraction................................................................................................................................... Factorial Design......................................................................................................................... CONCLUSION...........................................................................................................................ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationBACKGROUNDTherapeutically, Fluoxetine hydrochloride is a classified as a selective serotonin-reuptake inhibitor. Effectively used for the treatment of various depressions. Fluoxetine hydrochloride has been shown to have comparable efficacy to tricyclic antidepressants but with fewer anticholinergic side effects. The patent expiry becomes effective in 2001 (US). INTRODUCTIONFluoxetine capsules were prepared in two dosage strengths: 10mg and 20mg dosage strengths with the same capsule weight. The formulas are essentially similar and geometrically equivalent with the same ingredients and proportions. Minor changes in non-active proportions account for the change in active ingredient amounts from the 10 and 20 mg strength.The following validation, for the method SI-IAG-206-02 , includes assay and determination of Meta-Fluoxetine by HPLC, is based on the analytical method validation SI-IAG-209-06. Currently the method is the in-house method performed for Stability Studies. The Validation was performed on the 20mg dosage samples, IAG-21-001 and IAG-21-002.In the forced degradation studies, the two placebo samples were also used. PRECISIONSYSTEM REPEATABILITYFive replicate injections of the standard solution at the concentration of 0.4242mg/mL as described in method SI-IAG-206-02 were made and the relative standard deviation (RSD) of the peak areas was calculated.SAMPLE PEAK AREA#15390#25406#35405#45405#55406Average5402.7SD 6.1% RSD0.1ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::PRECISION - Method RepeatabilityThe full HPLC method as described in SI-IAG-206-02 was carried-out on the finished product IAG-21-001 for the 20mg dosage form. The method repeated six times and the relative standard deviation (RSD) was calculated.SAMPLENumber%ASSAYof labeled amountI 96.9II 97.8III 98.2IV 97.4V 97.7VI 98.5(%) Average97.7SD 0.6(%) RSD0.6PRECISION - Intermediate PrecisionThe full method as described in SI-IAG-206-02 was carried-out on the finished product IAG-21-001 for the 20mg dosage form. The method was repeated six times by a second analyst on a different day using a different HPLC instrument. The average assay and the relative standard deviation (RSD) were calculated.SAMPLENumber% ASSAYof labeled amountI 98.3II 96.3III 94.6IV 96.3V 97.8VI 93.3Average (%)96.1SD 2.0RSD (%)2.1The difference between the average results of method repeatability and the intermediate precision is 1.7%.HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationLINEARITYStandard solutions were prepared at 50% to 200% of the nominal concentration required by the assay procedure. Linear regression analysis demonstrated acceptability of the method for quantitative analysis over the concentration range required. Y-Intercept was found to be insignificant.RANGEDifferent concentrations of the sample (IAG-21-001) for the 20mg dosage form were prepared, covering between 50% - 200% of the nominal weight of the sample.Conc. (%)Conc. (mg/mL)Peak Area% Assayof labeled amount500.20116235096.7700.27935334099.21000.39734463296.61500.64480757797.52000.79448939497.9(%) Average97.6SD 1.0(%) RSD 1.0ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::RANGE (cont.)The results demonstrate linearity as well over the specified range.Correlation coefficient (RSQ)0.99981 Slope11808.3Y -Interceptresponse at 100%* 100 (%) 0.3%ACCURACYACCURACY OF STANDARD INJECTIONSFive (5) replicate injections of the working standard solution at concentration of 0.4242mg/mL, as described in method SI-IAG-206-02 were made.INJECTIONNO.PEAK AREA%ACCURACYI 539299.7II 540599.9III 540499.9IV 5406100.0V 5407100.0Average 5402.899.9%SD 6.10.1RSD, (%)0.10.1The percent deviation from the true value wasdetermined from the linear regression lineHPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::ACCURACY OF THE DRUG PRODUCTAdmixtures of non-actives (placebo, batch IAG-21-001 ) with Fluoxetine HCl were prepared at the same proportion as in a capsule (70%-180% of the nominal concentration).Three preparations were made for each concentration and the recovery was calculated.Conc.(%)Placebo Wt.(mg)Fluoxetine HCl Wt.(mg)Peak Area%Accuracy Average (%)70%7079.477.843465102.27079.687.873427100.77079.618.013465100.0101.0100%10079.6211.25476397.910080.8011.42491799.610079.6011.42485498.398.6130%13079.7214.90640599.413080.3114.75632899.213081.3314.766402100.399.618079.9920.10863699.318079.3820.45879499.418080.0820.32874899.599.4Placebo, Batch Lot IAG-21-001HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::VALIDATION OF FLUOXETINE HClAT LOW CONCENTRATIONLINEARITY AT LOW CONCENTRATIONSStandard solution of Fluoxetine were prepared at approximately 0.02%-1.0% of the working concentration required by the method SI-IAG-206-02. Linear regression analysis demonstrated acceptability of the method for quantitative analysis over this range.ACCURACY OF FLUOXETINE HCl AT LOW CONCENTRATIONThe peak areas of the standard solution at the working concentration were measured and the percent deviation from the true value, as determined from the linear regression was calculated.SAMPLECONC.µg/100mLAREA FOUND%ACCURACYI 470.56258499.7II 470.56359098.1III 470.561585101.3IV 470.561940100.7V 470.56252599.8VI 470.56271599.5(%) AverageSlope = 132.7395299.9SD Y-Intercept = -65.872371.1(%) RSD1.1HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationSystem RepeatabilitySix replicate injections of standard solution at 0.02% and 0.05% of working concentration as described in method SI-IAG-206-02 were made and the relative standard deviation was calculated.SAMPLE FLUOXETINE HCl AREA0.02%0.05%I10173623II11503731III10103475IV10623390V10393315VI10953235Average10623462RSD, (%) 5.0 5.4Quantitation Limit - QLThe quantitation limit ( QL) was established by determining the minimum level at which the analyte was quantified. The quantitation limit for Fluoxetine HCl is 0.02% of the working standard concentration with resulting RSD (for six injections) of 5.0%. Detection Limit - DLThe detection limit (DL) was established by determining the minimum level at which the analyte was reliably detected. The detection limit of Fluoxetine HCl is about 0.01% of the working standard concentration.ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::VALIDATION FOR META-FLUOXETINE HCl(EVALUATING POSSIBLE IMPURITIES)Meta-Fluoxetine HCl linearity at 0.05% - 1.0%Relative Response Factor (F)Relative response factor for Meta-Fluoxetine HCl was determined as slope of Fluoxetine HCl divided by the slope of Meta-Fluoxetine HCl from the linearity graphs (analysed at the same time).F =132.7395274.859534= 1.8Detection Limit (DL) for Fluoxetine HClThe detection limit (DL) was established by determining the minimum level at which the analyte was reliably detected.Detection limit for Meta Fluoxetine HCl is about 0.02%.Quantitation Limit (QL) for Meta-Fluoxetine HClThe QL is determined by the analysis of samples with known concentration of Meta-Fluoxetine HCl and by establishing the minimum level at which the Meta-Fluoxetine HCl can be quantified with acceptable accuracy and precision.Six individual preparations of standard and placebo spiked with Meta-Fluoxetine HCl solution to give solution with 0.05% of Meta Fluoxetine HCl, were injected into the HPLC and the recovery was calculated.HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::META-FLUOXETINE HCl[RECOVERY IN SPIKED SAMPLES].Approx.Conc.(%)Known Conc.(µg/100ml)Area in SpikedSampleFound Conc.(µg/100mL)Recovery (%)0.0521.783326125.735118.10.0521.783326825.821118.50.0521.783292021.55799.00.0521.783324125.490117.00.0521.783287220.96996.30.0521.783328526.030119.5(%) AVERAGE111.4SD The recovery result of 6 samples is between 80%-120%.10.7(%) RSDQL for Meta Fluoxetine HCl is 0.05%.9.6Accuracy for Meta Fluoxetine HClDetermination of Accuracy for Meta-Fluoxetine HCl impurity was assessed using triplicate samples (of the drug product) spiked with known quantities of Meta Fluoxetine HCl impurity at three concentrations levels (namely 80%, 100% and 120% of the specified limit - 0.05%).The results are within specifications:For 0.4% and 0.5% recovery of 85% -115%For 0.6% recovery of 90%-110%HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::META-FLUOXETINE HCl[RECOVERY IN SPIKED SAMPLES]Approx.Conc.(%)Known Conc.(µg/100mL)Area in spikedSample Found Conc.(µg/100mL)Recovery (%)[0.4%]0.4174.2614283182.66104.820.4174.2614606187.11107.370.4174.2614351183.59105.36[0.5%]0.5217.8317344224.85103.220.5217.8316713216.1599.230.5217.8317341224.81103.20[0.6%]0.6261.3918367238.9591.420.6261.3920606269.81103.220.6261.3920237264.73101.28RECOVERY DATA DETERMINED IN SPIKED SAMPLESHPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::REPEATABILITYMethod Repeatability - Meta Fluoxetine HClThe full method (as described in SI-IAG-206-02) was carried out on the finished drug product representing lot number IAG-21-001-(1). The HPLC method repeated serially, six times and the relative standard deviation (RSD) was calculated.IAG-21-001 20mg CAPSULES - FLUOXETINESample% Meta Fluoxetine % Meta-Fluoxetine 1 in Spiked Solution10.0260.09520.0270.08630.0320.07740.0300.07450.0240.09060.0280.063AVERAGE (%)0.0280.081SD 0.0030.012RSD, (%)10.314.51NOTE :All results are less than QL (0.05%) therefore spiked samples with 0.05% Meta Fluoxetine HCl were injected.HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationED. N0: 04Effective Date:APPROVED::Intermediate Precision - Meta-Fluoxetine HClThe full method as described in SI-IAG-206-02 was applied on the finished product IAG-21-001-(1) .It was repeated six times, with a different analyst on a different day using a different HPLC instrument.The difference between the average results obtained by the method repeatability and the intermediate precision was less than 30.0%, (11.4% for Meta-Fluoxetine HCl as is and 28.5% for spiked solution).IAG-21-001 20mg - CAPSULES FLUOXETINESample N o:Percentage Meta-fluoxetine% Meta-fluoxetine 1 in spiked solution10.0260.06920.0270.05730.0120.06140.0210.05850.0360.05560.0270.079(%) AVERAGE0.0250.063SD 0.0080.009(%) RSD31.514.51NOTE:All results obtained were well below the QL (0.05%) thus spiked samples slightly greater than 0.05% Meta-Fluoxetine HCl were injected. The RSD at the QL of the spiked solution was 14.5%HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationSPECIFICITY - STABILITY INDICATING EVALUATIONDemonstration of the Stability Indicating parameters of the HPLC assay method [SI-IAG-206-02] for Fluoxetine 10 & 20mg capsules, a suitable photo-diode array detector was incorporated utilizing a commercial chromatography software managing system2, and applied to analyze a range of stressed samples of the finished drug product.GLOSSARY of PEAK PURITY RESULT NOTATION (as reported2):Purity Angle-is a measure of spectral non-homogeneity across a peak, i.e. the weighed average of all spectral contrast angles calculated by comparing all spectra in the integrated peak against the peak apex spectrum.Purity Threshold-is the sum of noise angle3 and solvent angle4. It is the limit of detection of shape differences between two spectra.Match Angle-is a comparison of the spectrum at the peak apex against a library spectrum.Match Threshold-is the sum of the match noise angle3 and match solvent angle4.3Noise Angle-is a measure of spectral non-homogeneity caused by system noise.4Solvent Angle-is a measure of spectral non-homogeneity caused by solvent composition.OVERVIEWT he assay of the main peak in each stressed solution is calculated according to the assay method SI-IAG-206-02, against the Standard Solution, injected on the same day.I f the Purity Angle is smaller than the Purity Threshold and the Match Angle is smaller than the Match Threshold, no significant differences between spectra can be detected. As a result no spectroscopic evidence for co-elution is evident and the peak is considered to be pure.T he stressed condition study indicated that the Fluoxetine peak is free from any appreciable degradation interference under the stressed conditions tested. Observed degradation products peaks were well separated from the main peak.1® PDA-996 Waters™ ; 2[Millennium 2010]ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationFORCED DEGRADATION OF FINISHED PRODUCT & STANDARD 1.UNSTRESSED SAMPLE1.1.Sample IAG-21-001 (2) (20mg/capsule) was prepared as stated in SI-IAG-206-02 and injected into the HPLC system. The calculated assay is 98.5%.SAMPLE - UNSTRESSEDFluoxetine:Purity Angle:0.075Match Angle:0.407Purity Threshold:0.142Match Threshold:0.4251.2.Standard solution was prepared as stated in method SI-IAG-206-02 and injected into the HPLC system. The calculated assay is 100.0%.Fluoxetine:Purity Angle:0.078Match Angle:0.379Purity Threshold:0.146Match Threshold:0.4272.ACID HYDROLYSIS2.1.Sample solution of IAG-21-001 (2) (20mg/capsule) was prepared as in method SI-IAG-206-02 : An amount equivalent to 20mg Fluoxetine was weighed into a 50mL volumetric flask. 20mL Diluent was added and the solution sonicated for 10 minutes. 1mL of conc. HCl was added to this solution The solution was allowed to stand for 18 hours, then adjusted to about pH = 5.5 with NaOH 10N, made up to volume with Diluent and injected into the HPLC system after filtration.Fluoxetine peak intensity did NOT decrease. Assay result obtained - 98.8%.SAMPLE- ACID HYDROLYSISFluoxetine peak:Purity Angle:0.055Match Angle:0.143Purity Threshold:0.096Match Threshold:0.3712.2.Standard solution was prepared as in method SI-IAG-206-02 : about 22mg Fluoxetine HCl were weighed into a 50mL volumetric flask. 20mL Diluent were added. 2mL of conc. HCl were added to this solution. The solution was allowed to stand for 18 hours, then adjusted to about pH = 5.5 with NaOH 10N, made up to volume with Diluent and injected into the HPLC system.Fluoxetine peak intensity did NOT decrease. Assay result obtained - 97.2%.ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationSTANDARD - ACID HYDROLYSISFluoxetine peak:Purity Angle:0.060Match Angle:0.060Purity Threshold:0.099Match Threshold:0.3713.BASE HYDROLYSIS3.1.Sample solution of IAG-21-001 (2) (20mg/capsule) was prepared as per method SI-IAG-206-02 : An amount equivalent to 20mg Fluoxetine was weight into a 50mL volumetric flask. 20mL Diluent was added and the solution sonicated for 10 minutes. 1mL of 5N NaOH was added to this solution. The solution was allowed to stand for 18 hours, then adjusted to about pH = 5.5 with 5N HCl, made up to volume with Diluent and injected into the HPLC system.Fluoxetine peak intensity did NOT decrease. Assay result obtained - 99.3%.SAMPLE - BASE HYDROLYSISFluoxetine peak:Purity Angle:0.063Match Angle:0.065Purity Threshold:0.099Match Threshold:0.3623.2.Standard stock solution was prepared as per method SI-IAG-206-02 : About 22mg Fluoxetine HCl was weighed into a 50mL volumetric flask. 20mL Diluent was added. 2mL of 5N NaOH was added to this solution. The solution was allowed to stand for 18 hours, then adjusted to about pH=5.5 with 5N HCl, made up to volume with Diluent and injected into the HPLC system.Fluoxetine peak intensity did NOT decrease - 99.5%.STANDARD - BASE HYDROLYSISFluoxetine peak:Purity Angle:0.081Match Angle:0.096Purity Threshold:0.103Match Threshold:0.3634.OXIDATION4.1.Sample solution of IAG-21-001 (2) (20mg/capsule) was prepared as per method SI-IAG-206-02. An equivalent to 20mg Fluoxetine was weighed into a 50mL volumetric flask. 20mL Diluent added and the solution sonicated for 10 minutes.1.0mL of 30% H2O2 was added to the solution and allowed to stand for 5 hours, then made up to volume with Diluent, filtered and injected into HPLC system.Fluoxetine peak intensity decreased to 95.2%.ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationSAMPLE - OXIDATIONFluoxetine peak:Purity Angle:0.090Match Angle:0.400Purity Threshold:0.154Match Threshold:0.4294.2.Standard solution was prepared as in method SI-IAG-206-02 : about 22mg Fluoxetine HCl were weighed into a 50mL volumetric flask and 25mL Diluent were added. 2mL of 30% H2O2 were added to this solution which was standing for 5 hours, made up to volume with Diluent and injected into the HPLC system.Fluoxetine peak intensity decreased to 95.8%.STANDARD - OXIDATIONFluoxetine peak:Purity Angle:0.083Match Angle:0.416Purity Threshold:0.153Match Threshold:0.4295.SUNLIGHT5.1.Sample solution of IAG-21-001 (2) (20mg/capsule) was prepared as in method SI-IAG-206-02 . The solution was exposed to 500w/hr. cell sunlight for 1hour. The BST was set to 35°C and the ACT was 45°C. The vials were placed in a horizontal position (4mm vials, National + Septum were used). A Dark control solution was tested. A 2%w/v quinine solution was used as the reference absorbance solution.Fluoxetine peak decreased to 91.2% and the dark control solution showed assay of 97.0%. The difference in the absorbance in the quinine solution is 0.4227AU.Additional peak was observed at RRT of 1.5 (2.7%).The total percent of Fluoxetine peak with the degradation peak is about 93.9%.SAMPLE - SUNLIGHTFluoxetine peak:Purity Angle:0.093Match Angle:0.583Purity Threshold:0.148Match Threshold:0.825 ED. N0: 04Effective Date:APPROVED::HPLC ASSAY with DETERMINATION OF META-FLUOXETINE HCl.ANALYTICAL METHOD VALIDATION10 and 20mg Fluoxetine Capsules HPLC DeterminationSUNLIGHT (Cont.)5.2.Working standard solution was prepared as in method SI-IAG-206-02 . The solution was exposed to 500w/hr. cell sunlight for 1.5 hour. The BST was set to 35°C and the ACT was 42°C. The vials were placed in a horizontal position (4mm vials, National + Septum were used). A Dark control solution was tested. A 2%w/v quinine solution was used as the reference absorbance solution.Fluoxetine peak was decreased to 95.2% and the dark control solution showed assay of 99.5%.The difference in the absorbance in the quinine solution is 0.4227AU.Additional peak were observed at RRT of 1.5 (2.3).The total percent of Fluoxetine peak with the degradation peak is about 97.5%. STANDARD - SUNLIGHTFluoxetine peak:Purity Angle:0.067Match Angle:0.389Purity Threshold:0.134Match Threshold:0.8196.HEAT OF SOLUTION6.1.Sample solution of IAG-21-001-(2) (20 mg/capsule) was prepared as in method SI-IAG-206-02 . Equivalent to 20mg Fluoxetine was weighed into a 50mL volumetric flask. 20mL Diluent was added and the solution was sonicated for 10 minutes and made up to volume with Diluent. 4mL solution was transferred into a suitable crucible, heated at 105°C in an oven for 2 hours. The sample was cooled to ambient temperature, filtered and injected into the HPLC system.Fluoxetine peak was decreased to 93.3%.SAMPLE - HEAT OF SOLUTION [105o C]Fluoxetine peak:Purity Angle:0.062Match Angle:0.460Purity Threshold:0.131Match Threshold:0.8186.2.Standard Working Solution (WS) was prepared under method SI-IAG-206-02 . 4mL of the working solution was transferred into a suitable crucible, placed in an oven at 105°C for 2 hours, cooled to ambient temperature and injected into the HPLC system.Fluoxetine peak intensity did not decrease - 100.5%.ED. N0: 04Effective Date:APPROVED::。
基于SIF模式特性的曼彻斯特改进算法李丞;张玉;唐波;刘垒【摘要】针对解决二次监视雷达(SSR)应答信号基于曼彻斯特编码特性分选时,会出现因混扰信号中包含SIF模式信号而不能完全分选,在基于曼彻斯特算法的基础上提出改进算法.该算法首先根据信号模式特征之间的不同,对混扰信号进行针对性预处理.然后通过求解得出分离矩阵,以实现对混扰信号中的SIF模式部分进行有效分选.最后对非SIF模式类信号进行分选.仿真验证表明,该改进算法在低信噪比时可以有效进行混模式信号分选,且其性能在相较同类算法时具有更好的抗噪性能与分选性能.%To overcome decoding errors problem of secondary surveillance radar (SSR) due to the phenomenon of different modes signals including SIF mode signals garbling,an improved algorithm based on the Manchester encoding property of mode S is proposed for separating the garbled replies. Firstly,the garbling signals are pre-conducting based on the features of different modes.Then a method is proposed to solve for the separation matrix,and the SIF signals in garbled signals can be de-garbled .Last the other signals can be de-garbled too.Numerical simulations show that the improved algorithm maintains good separation performance in the low SNR case,and its performance is better in separation and anti-noise when in the same time delay than same algorithms.【期刊名称】《火力与指挥控制》【年(卷),期】2018(043)004【总页数】6页(P42-47)【关键词】二次雷达;信号分选;曼彻斯特编码;SIF模式【作者】李丞;张玉;唐波;刘垒【作者单位】国防科技大学电子对抗学院,合肥230037;国防科技大学电子对抗学院,合肥230037;国防科技大学电子对抗学院,合肥230037;国防科技大学电子对抗学院,合肥230037【正文语种】中文【中图分类】TN958.960 引言二次雷达系统（Second Surveillance Radar，SSR）发展到今天，已经成为空中交通管制系统的基础［1］，其系统中的信号模式也随着发展更新换代。
胆囊癌临床诊疗的新进展中华外科杂志普外空间 2022-08-10 10:00 发表于北京作者：杨自逸,刘诗蕾,蔡晨,吴自友,熊逸晨,李茂岚,吴向嵩,全志伟,龚伟文章来源：中华外科杂志, 2022, 60(8)摘要胆囊癌的恶性程度极高，尚缺乏早期诊断方法和有效治疗手段，亟需高质量研究突破诊疗瓶颈。
本文回顾了2021年国内外发表的胆囊癌研究相关文献，对临床诊疗领域的重要进展进行综述，详细介绍了胆囊癌最新流行病学数据及危险因素、新兴的外周血实验室检查和影像学诊断方法、病理学类型新分类、外科治疗的热点与争议及系统性综合治疗动态。
这些研究结果有助于探索更有效的胆囊癌诊治方法，为改善胆囊癌患者的预后带来希望。
胆囊癌是胆道系统常见的恶性肿瘤，具有症状隐匿、发展迅速、早期转移、预后极差的特点。
我国是胆囊癌的高发地区之一，近年来发病率和病死率呈缓慢上升趋势。
目前仍缺乏特异度和灵敏度均较好的胆囊癌早期诊断手段，临床发现的胆囊癌多为中晚期。
尽管医学科技不断发展，早期诊断和根治性手术切除仍是可能治愈胆囊癌的手段，行之有效的系统性治疗方法依然在不断探索中。
本文展示了2021年胆囊癌临床诊疗领域的研究进展，以探索更好的胆囊癌诊疗方法。
一、流行病学特征（一）发病率与死亡率2020年全球癌症统计数据显示，全球胆囊癌新发115 949例（男性41 062例，女性74 887例），死亡84 695例（男性30 265例，女性54 430例）［1］，均居消化系统肿瘤第6位。
胆囊癌全球发病率存在明显的地域差异，全球年龄标准化发病率平均为2.3/10万人，以东亚、南美最高，西欧、北美则发病率较低［2］；且近年来男性和年轻群体的胆囊癌发病率呈升高趋势。
我国国家癌症中心数据显示，国内胆囊癌发病率为3.95/10万人（男性3.70/10万人，女性4.21/10万人），死亡率为2.95/10万人（男性1.9/10万人，女性2.1/10万人）［3］。
日本血吸虫组织蛋白酶B基因的克隆、表达及初步
鉴定的开题报告
一、背景和研究意义
日本血吸虫是人们非常关注的寄生虫之一，造成了许多人类疾病。
组织蛋白酶B是日本血吸虫内切蛋白酶家族的成员之一，起着非常重要
的作用，可以降解寄主体内的组织蛋白，从而破坏宿主免疫系统，并保
证寄生虫的生存和繁殖。
因此，对于组织蛋白酶B基因的研究，可以深
入了解日本血吸虫在宿主体内的生存和致病机制，为研发相关的治疗和
预防措施提供重要的基础知识。
二、研究内容和方法
本研究通过PCR扩增和克隆技术，成功地获得了日本血吸虫组织蛋
白酶B基因的全长序列，并将其构建到表达载体中。
然后，采用大肠杆
菌系统表达目标蛋白，并通过蛋白质纯化技术得到高纯度的重组蛋白。
最后，利用西方印迹等方法对目标蛋白进行了初步鉴定。
三、预期结果
本研究预计得到完整的日本血吸虫组织蛋白酶B基因序列，同时也
可以得到高水平的表达和纯化的目标蛋白。
通过初步的鉴定和分析，可
以初步了解该蛋白的结构和功能等方面的特点，为日后深入研究和开发
相关药物提供必要的前提和基础。
四、研究意义和应用前景
该研究结果对于深入了解日本血吸虫在宿主体内的生存和繁殖机制，以及研发相关药物具有重要意义。
同时，本研究也为其他相关寄生虫的
研究提供了新的思路和方法，并对于巩固我国在相关研究领域的学术地
位和国际地位具有积极的推动和作用。
《别样的英雄》(A Special Kind of Hero) 演唱者：斯黛芬妮·劳伦斯(Stephanie Lawrence)1990年意大利世界杯:《意大利之夏》(UN’ESTATE ITALIANA)英语版本名称为：To Be Number One 演唱者：吉奥吉.莫罗德(Giorgio Moroder)和吉娜.娜尼尼(Gianna Nannini)“意大利之夏”或许是最成功的世界杯主题曲，至今仍被资深球迷和歌迷所津津乐道。
1994年美国世界杯:“荣耀之地”(Gloryland) 演唱者：达利尔.豪(Daryl Hall)美国一直是足球运动的处女地，把世界杯主办权交到他们手里似乎是个错误的选择，主题曲也变得黯淡。
1998法国世界杯:从1998年起，世界杯赛的主题曲不再仅限于一首，而且开始灌录世界杯官方专辑唱片。
1998年的《Allez! Ola! Ole! 》中就收录了15首代表参赛各国的足球歌曲。
官方主题歌为“我踢球你介意吗”和“生命之杯”。
《我踢球你介意吗》(法语：La Cour des Grands，英语：Do you mind if I play)(1998年法国世界杯主题曲1) 演唱者：尤索·恩多(Youssou N’Dour )& 阿克塞拉·瑞德(Axelle Red) “我踢球你介意吗”是首轻快的歌曲，带着浓烈的热带情调和欢快的吟唱风格。
〈生命之杯〉(La Copa De La Vida) 演唱者：瑞奇.马汀(Ricky Martin) “生命之杯” 在世界杯之后也传播甚广，成为很多足球节目用来烘托气氛的第一选用曲目。
歌曲中的鼓乐节奏和号角奏鸣都颇为煽情。
2002年日韩世界杯:《风暴》(boom) 演唱者：阿纳斯塔西娅(Anastacia) “风暴”曲调简洁、节奏强劲，流行的曲风给人异域感觉，和“生命之杯” 相比它少了些火般热情，多了份紧迫感强劲的冲击。
玛咖大事件1961年……生物学家Gloria Chacon首次公布玛卡的研究成果得到业界的极大关注。
1992年……联合国粮农组织罗马宣言将玛咖作为一种难得的营养补充剂推向世界。
1992年……国际植物遗传研究所将玛咖列为珍惜植物。
1998年……美洲植物疗法研究所玛咖专注《玛咖秘鲁要用和营养植物》。
1999年……美国科学家发现玛咖中含有两种新的活性成分，玛咖酰胺和玛咖烯，并确定这两种物质对平衡人体荷尔蒙分泌有显著作用，所以玛咖又被称为天然荷尔蒙发动机。
2000年……国际登山组织联盟指定产品。
2001年……美国食品与药品管理局通过了玛咖保健品进入美国的认证。
2001年……美国专利局批准第一项从玛咖中提取的有效成分专利。
2001年……美国为玛咖中发现的两种新物质命名为玛咖酰胺和玛咖烯。
2001年……美国太空总署首次将玛咖作为宇航员的太空粮食。
2001年……路透社报道玛咖的神奇功效。
2001年……中央电视台“新闻30分”专题报道玛咖最新研究成果。
2001年……中国兴奋剂及营养测试研究中心确认玛咖不含国际奥委会禁用成分。
2001年……中国工程院院士肖培根《玛咖—全世界瞩目的保健食品》一文发表。
2001年……中国云南玛咖引种成功。
2002年……赵言林首次将玛咖与国内草本植物结合，对自己进行康复体验，获得显著效果。
2002年……玛咖成为韩日世界杯运动员指定营养品。
2002年……中国卫生部正式批准玛咖进入中国。
2003年……华中科技大学余江龙教授《国际良种—药食两用植物玛咖》一书出版。
2004年……赵言林在MACA与草本植物相结合的应用效果方面，展开长期观察和探索。
2007年……赵言林撰写的论文《中外草药在治未病和已病领域的探索与发现》在中华中西医临床杂志首次发表，玛咖打破传统应用束缚与中国草本植物相结合迈出了第一步。
2011年……我国卫生部第13号公告批准玛卡粉作为新资源食品。
2011年……中国第一款玛咖精片——巨青玛咖玛咖精片诞生。
综述以DNA拓扑异构酶Ò为靶点的抗癌药物蒙凌华,张永炜,丁健(中国科学院上海药物研究所肿瘤药理实验室,上海200031)[摘要]DNA拓扑异构酶Ò(T opoisomeraseÒ,T OPOÒ)是一种真核生物生存所必需的泛酶,在几乎所有DNA代谢过程中发挥重要作用。
T OPOÒ使一条完整的DNA双链穿过一个移过性的双链断口,从而导致DNA解结或解旋。
因为T OPOÒ具有重要的生理功能,它已成为抗癌药物的重要作用靶点。
以T OPOÒ为靶点的药物按作用方式可分为2类:一类通过稳定T O POÒ介导的可切割复合物而杀死肿瘤细胞,称为T OPOÒ毒剂(T OP OÒpoison);另一类通过抑制T O POÒ的催化活性而达到抑制肿瘤的作用,称为T OPOÒ催化抑制剂(T O POÒinhib-i tor)。
近年来,对T OPOÒ催化机制和药物作用方式的研究取得了很大进展,这些发现有助于进一步了解T OPO Ò的生理功能,进而研究出更有效的治疗方案和新的抗癌药。
本文介绍了以T OPOÒ为靶点的抗癌药物的作用机制及其发展现状。
[关键词]抗癌药物;拓扑异构酶Ò;肿瘤[中图分类号]R973[文献标识码]A[文章编号]1003-3734(2002)09-0675-09Anticancer drugs targeting DNA topoisomeraseÒMENG Ling-hua,ZHANG Yong-wei,DING Jian(Div ision of A ntitumor Phar macology,Shanghai I nstitute o f Materia M edica,Chinese Academy o f Sciences,Shanghai200031,China)[Abstract]DNA topoisom eraseÒ(T OPOÒ)is a ubiquitous enzym e that is essential for sur-vival of the eukaryotic org anisms,playing an important role in most processes of the DNA mechanisms. T he enzyme unknots or decatenates DNA by passing an intact helix through a transient double-strand-ed break.In addition to its critical cellular functions,TOPOÒis an important target for a number of the most active and w idely prescribed anticancer drugs.Despite the fact that they share the same tar-get,drugs targ eting TOPOÒhave different modes of action:TOPOÒpoisons kill cell by stabilizing the DNA-TOPOÒcleavable complex;w hile TOPOÒcatalytic inhibitors act by blocking overall cata-lytic activity.Over the past several years,great prog ress has been made in the cataly tic mechanisms of T OPOÒand the mechanism of action of drugs targeting this enzyme.These advances have provided novel insight into the physiological functions of TOPOÒand have led to the development of more eff-i cacious chemotherapeutic reg imens and new anticancer drugs.The ant-i cancer mechanism of TOPOÒinhibitors and the status of development are introduced in this review.[Key words]anticancer drug;topoisomeraseÒ;tumor脱氧核糖核酸(DNA)是绝大多数生物的遗传物质,具有稳定、多样和能够自我复制的特点。
三氧化二砷对人乳腺癌雌激素受体表达的影响的开题报告一、研究背景人乳腺癌是一种常见的恶性肿瘤，其中雌激素受体（ER）是诱发乳腺癌的主要因素之一。
三氧化二砷（As2O3）是一种广泛使用的抗肿瘤药物，具有诱导癌细胞凋亡、抑制细胞增殖和转移等作用。
虽然As2O3已被广泛应用于肿瘤治疗，但其对乳腺癌ER表达水平的影响尚不完全清楚。
因此，探究As2O3对乳腺癌ER表达的影响，有助于为乳腺癌的诊断和治疗提供新的思路和方法。
二、研究目的本研究旨在探究As2O3对人乳腺癌ER表达的影响，为乳腺癌的早期诊断和治疗提供新的实验数据和思路。
三、研究内容1. 收集人乳腺癌组织样本，并分离细胞；2. 通过Western blot方法检测乳腺癌组织样本中ER的表达水平；3. 将细胞分为对照组和As2O3处理组，处理组分别在不同浓度下进行处理；4. 对处理组进行Western blot检测，探究As2O3处理对ER表达的影响；5. 通过统计分析方法，对实验数据进行统计和分析。
四、研究意义1. 探究As2O3对乳腺癌ER表达的影响，有助于为乳腺癌的早期诊断和治疗提供新的思路和方法；2. 为深入探究As2O3的药理作用提供新的实验数据；3. 为乳腺癌治疗提供新的治疗方法和治疗方案。
五、研究方法1. 收集人乳腺癌组织样本，并从中分离细胞；2. Western blot方法检测乳腺癌组织样本中ER的表达水平；3. 将细胞分为对照组和As2O3处理组，处理组分别在不同浓度下进行处理；4. Western blot方法检测乳腺癌细胞As2O3处理后ER的表达水平；5. 对实验数据进行统计和分析。
六、研究预期结果1. 通过Western blot方法，检测出乳腺癌组织中ER的表达水平；2. 在不同浓度下处理乳腺癌细胞，并通过Western blot检测出As2O3处理后ER的表达水平；3. 分析实验数据，探究As2O3处理对乳腺癌ER表达的影响。
An adaptive scheduling service for Real-Time CORBAAlexandre Cervieri1, Rômulo Silva de Oliveira2, Cláudio F. Resin Geyer11Universidade Federal do Rio Grande do Sul - UFRGSInstituto de Informática - Av. Bento Gonçalves, 9500 – Bloco IVPorto Alegre, RS – Brasil{cervieri, geyer}@inf.ufrgs.br2Departamento de Automação e Sistemas - DAS-CTC-UFSCCaixa Postal 476 - CEP 88040-900Florianópolis, SC – Brasilromulo@das.ufsc.brAbstract. CORBA is an important standard middleware used in thedevelopment of distributed applications. It has also been used with distributedreal-time applications, through its extension for real-time systems, RT-CORBA.RT-CORBA includes many mechanisms to reduce the non-determinismassociated with ordinary CORBA. These mechanisms can be used to provideguarantees for hard real-time systems if the right support from the operatingsystem and network protocols is available. RT-CORBA mechanisms can alsobe used to improve the timing behavior of soft real-time applications, when thelower layers are not able to provide guarantees. This paper proposes an adaptivescheduling service in the context of RT-CORBA to support the implementationof distributed soft real-time applications. The proposal is based on theadaptation of task periods, so as to reduce system load while still trying to meetthe original deadline of all tasks. This is a best-effort approach that dynamicallyprovides graceful degradation in case of overload. The adaptive serviceproposed in this paper is validated by a set of experiences based on mechanismsof RT-CORBA and TAO, the ORB implementation used.1 IntroductionThe large-scale availability of computer networks has motivated the search for ways to facilitate and to accelerate the development of applications in distributed systems. CORBA (Common Object Request Broker Architecture), created as an initiative of OMG (Object Management Group), is a middleware that provides a high degree of language and platform independence for the application [1] [2].However, CORBA at first had no mechanisms to define and guarantee temporal behavior, as well as QoS requirements. These initial limitations made CORBA not appropriate for real-time applications. In order to overcome these deficiencies, OMG started an effort to define a real-time extension to CORBA denominated RT-CORBA [3] [4].RT-CORBA is still a standard under development. It includes interfaces and mechanisms to allow the definition and the execution of a great variety of real-timeapplications. Even so, many of the mechanisms that try to maintain the predictability of the applications depend on another aspects such as support from the operating system to guarantee time requirements, a more predictable communication system, scheduling policies and mechanisms that guarantee the execution of tasks at the correct moment [5].RT-CORBA still has many opened questions on this latter aspect. One of the essential points for the development of a real-time system is the choice of the correct scheduling policy for a given class of applications. Hard real-time applications do not allow deadline misses. Consequently, they demand a scheduling policy capable of providing off-line guarantees. Soft real-time applications allow a run-time scheduling approach, by the use of best-effort policies. In this case some quality degradation is accepted in order to satisfy deadlines, and even delays are allowed to same degree.The aim of this work is to present an adaptive scheduling service in the context of RT-CORBA to support the implementation of distributed real-time applications. The proposal is based on the adaptation of task periods, with the objective of minimizing system load while still trying to satisfy the deadlines of all tasks. The adaptive service proposed in this paper is validated by a set of experiences based on mechanisms of the RT-CORBA standard and of TAO, the specific ORB implementation used.This paper is organized so that in section 2 we make a brief revision of standard CORBA and RT-CORBA besides presenting some characteristics of TAO, the implementation used. Section 3 is a small revision of the literature on adaptive mechanisms in real-time systems. Finally, section 4 describes the adaptive scheduling service proposed. Its implementation is presented in section 5 and section 6 shows the results of the experiences. Final remarks appear in section 7.2 RT – CORBADefined by OMG, CORBA is nowadays one of the most complete architectures for the development of distributed systems applications. Many of the benefits associated with CORBA stand in the use of an independent language (IDL – Interface Definition Language) to define object interfaces. It is possible to integrate applications not written for distributed systems and event legacy software in these systems because of the existence of IDL mappings for several languages used in the market.The OMA (Object Management Architecture) is defined as a reference model for the technology of objects. The development of that architecture followed certain technical objectives so that it brought benefits to the management of objects, such as: conformity, distribution transparency, good performance in remote and local operations, expandable and dynamic behavior, architecture based on name service, access control, concurrency control, among others. OMG (Object Management Group) used the OMA to describe a generalization, a more abstract model, that is to say, a model of higher level, which is implemented by CORBA. It can be said that OMA serves as a “type” and CORBA as an “instance” of that type [6].The ORB (Object Request Broker) is the communication middleware that allows client objects to send request and receive replies from server objects, which can execute locally or remotely in relation to the client. They have defined mappings ofIDL to several languages to allow the access to the ORB, while not defining details about its implementation. That guarantees some freedom for the developers and assures that a program specified for an ORB will be partially portable to any other ORB that is compliant with CORBA definitions. The Object Adapter, for its time, has the function of controlling the life cycle of the servers. It is responsible for server creation, activation and destruction [7].Clients and servers can also access ORB services and the CORBA standard services defined by OMG. We can highlight, for its importance in this work: (i) the name service, whose function is to identify an object based on its name, returning its reference, and vice-versa; (ii) the event service, which defines a generic interface for sending messages between multiple sources and multiple destinations. The event service includes an event channel that allows the generation of events without consumers and suppliers having to communicate directly. Two forms of interaction are offered: (i) push, the supplier is active and sends events to the channel; (ii) pull, the consumer is active and tries to receive an event from the channel (this operation may be blocking or not) [1].In spite of the advantages that it brings to the development of distributed applications, CORBA 2.x does not satisfy all the demands of high performance and real-time applications, mainly because of the following reasons [8] [9]: lack of interfaces to define QoS (Quality of Service), lack of guarantees of QoS, lack of programming characteristics suitable for real-time and lack of optimizations for high performance.Due to the CORBA drawbacks in supporting real-time applications, OMG have created in 1995 a work group with the objective of extending CORBA specifications. In October 1998, the five proposals presented were united in a single document [8], RT-CORBA (included in the specification of CORBA 2.4). It is nowadays in a process of developing version 2, which will integrate CORBA 3.RT-CORBA identifies capacities that can be vertically (from the network layer to the application layer and vice-versa) and horizontally (end to end) integrated and managed by an ORB to guarantee the predictability about activities between CORBA clients and servers. Some of those capacities are:•Management of the communication infrastructure: RT-CORBA should provide polices and mechanisms for the management of the communication infrastructure, from the choice of a connection to the exploration of advanced QoS characteristics, including threads interface, thread pools, explicit binding and others; •Operating system scheduling mechanisms: ORBs explore the mechanisms of the operating system to schedule activities at the application level. RT-CORBA considers mainly real-time systems based on fixed priorities. In that way, these mechanisms correspond to the management of priorities of the threads scheduled by the operating system;•Real-time ORB: a real-time ORB should provide standardized interfaces to allow applications to specify their demands of resources to the ORB, besides supporting communication between clients and servers in a transparent way;•Real-time services and applications: a real-time ORB should also create an efficient environment, with end-to-end predictability, for high-level services and applications.Since OMG does not supply an implementation of CORBA and RT-CORBA, there are alternatives to be analyzed. Among CORBA implementations and, more specifically, RT-CORBA implementations, TAO (The ACE ORB) is an important one. TAO can be considered more than just an ORB to support real-time communication, it is a complete architecture for execution of real-time applications, with as much hard as soft deadlines. The architecture of TAO contains the following characteristics and elements [10]: a real-time I/O sub-system, RT-ORB, protocol GIOP for real-time, RT-POA, optimized IDL compiler and presentation layer, optimizations of the memory management by minimizing the copy of data and means for QoS specification.TAO is one of the ORBs that presents the most complete implementation of RT-CORBA. However, its implementation of some services differs from the standard definitions. One of those services is scheduling. The TAO scheduling service is responsible for allocating system resources in order to guarantee the needs of the applications. It was designed for hard real-time applications and its main objective is to guarantee that the demands for resources will be satisfied. It is divided in an off-line component and a run-time component. First, the possibility of a correct scheduling of all the operations is analyzed and their priorities are assigned. Run-time components offer fast access to priorities values and coordinate mode change operations.Another TAO service that was improved for the use in real-time applications was the event service. Basically the method of interaction push was affected [11]: (i) consumers and suppliers of events can specify their demand and execution characteristics using QoS parameters; (ii) correlation and filtering mechanisms are centralized in the event channel; (iii) consumers can specify time-out dependent events.3 Adaptation Mechanisms for Real-Time ApplicationsEven in computer systems specifically built for the execution of real-time applications there are many opportunities for adaptation. This can happen because of the load generated by the application not having a well-known limit that can be analyzed off-line. Even if the load has a well characterized demand for resources, it may not be economically feasible to guarantee its behavior in a worst-case scenario.Although most of the real-time literature is about adaptation through quality of service negotiation, adaptation can also be a unilateral action. In the case of a unilateral adaptation the following scenarios are possible:• A variation in the application behavior triggers an adaptation of the support.• A variation in the support behavior triggers an adaptation of the support itself.• A variation in the application triggers an adaptation of the application itself. In systems where the application can reserve resources, the application itself must manage reserved resources.• A variation in the support behavior triggers an adaptation of the application. This situation is very common in distributed systems, where variations of the response time can be associated with sending messages on the network.Real-time applications in a distributed environment are subject to variations in the response time of its tasks. Those variations can be caused by delays in message transmissions on the network or changes in some processing node used by the application. There is the need for a continuous adaptation during the application lifetime. Adaptation mechanisms described in the literature that can be used in this context will be examined in this section:• Delaying a task . The simplest and more frequent form of adaptation is simply to relax the deadline concept. One of the first papers proposing this approach appears in Jensen et. al. [12]. They propose that the conclusion of each task contributes to the system with a benefit and the value of this benefit can be expressed as a function of the instant of task conclusion (time-value function).• Changes in the task period . In a real-time application a lot of tasks are executed periodically. In general, these tasks have their period defined off-line. A way to provide adaptability in the application is to allow this period to vary dynamically during its execution. This way, the quality of the application, represented here by the period of its tasks, would be adapted to the performance of the platform where it executes.• Canceling a task execution . A more radical form of flexibilization is simply not to execute some tasks when the performance is below the desired level. In the case of applications with repetitive tasks, it is possible to cancel a specific task activation or to cancel the task completely.• Changes in the task execution time . In this adaptation mechanism, tasks are scheduled so they respect their respective deadlines. In case of overload, the task execution time is reduced. In order to implement that, it is necessary for each task to have options of the type quality versus execution time. This approach is usually known as Imprecise Computation [13].There are several proposals in the literature about how to measure system quality and how to act on the system to maximize its quality. For example, in Lu et. al. [14] the quality of a system is measured by the deadline miss rate and by the rate of system utilization in a given measuring window. Control is implemented by changing the operation mode of tasks and by deciding to accept or reject tasks for execution. Beccari et. al. [15] also identifies overloads in the system through the processor utilization and the control is implemented through graceful degradation of task periods. Other works found in the literature that follow the same direction can be found in Welch et. al. [16], Shin & Meissner [17] and Abdelzaher et. al. [18] [19].Table 1 presents a qualitative analysis of some adaptive approaches in the literature.Table 1. Some adaptive techniques in the literatureProposal Measuring Acting Experiences PlatformLu et. al. [14] Deadline missrate and system usage within a measuringwindow.Changes in the task operation mode andtake the decision whenaccept or reject new tasks. Simulation Single processor. Brandt et. System usage. CPU usage determines Linux/UNIX Singleal. [15] the system operation mode (requirementsand relative benefit). systems. processor.al. [20] of soft real time tasks (hard real time tasks are not affected),graceful degradation. and prototype still underdevelopment for a VME-based system.processor.Shin &Meissner [17]System quality (each task has a benefit value depending onthe period).Changes in the period of tasks and/or tasks relocation. Simulation (random tasks and sample application) Multi processor.et. al. [19] packets persecond.level (communication) required by the tasks. Group (TOG) 7.2 kernel (comm. system) Abdelzager et. al. [18] The system benefit is calculated when every new task arrives.Accept or reject the execution of tasksmodules or the whole task, following theQoS level negotiated. RTPOOLimplemented over OSFMach RT-mk7.2Multi processor.4 Adaptive Scheduling ServiceIn the previous section many adaptation mechanisms were introduced. Considering that a lot of real-time applications have its structure based on periodic activities, we opted for an approach directed to this type of activity. The proposal of this work was inspired by Lu et. al. [14] and other works that use a feedback loop to control the system during overloads.In the same way of Lu et. al. [14], our control of the system is based on information collected from all activities that compose the system. This information is analyzed by a periodic process, which will be called as AdaptiveService in this work. Lu et. al. [14] includes adaptation in the system by considering the utilization rate of the system (U(t)) and the deadline miss rate (MR(t), miss ratio function) of the tasks inside a measuring window (MW miss-ratio window). The measure of deadline miss-ratio is quite adapted for systems with firm deadlines, where there is no benefit in concluding a task after its deadline.Unlike the approach used by [14], which was created for a monoprocessor, we deal with the execution of soft real-time tasks on a distributed system. We decided to quantify the system quality by comparing the deadlines and the response times of the real-time tasks. This quality measure will be denominated delay profile (DY(t)), which is observed within a measuring window called delay window (DW). In thiswork it is assumed that the deadline of each task is equal to the nominal period of the same task (D=P). Therefore, DY(t) can be expressed as:∑−−=t DWt i i N D R t DY )()( t : current instant of time DW : delay window, the time window used for calculation of DY(t) R i : response time of task i for a certain activation D i : task deadlineN : number of conclusions (of all tasks)within window DW (1) Actuation is done on the task periods. This actuation could be made for each taskin a specific way. For example, there could be an adaptation factor especially calculated for each task by considering each task importance. In this work it is assumed that all tasks have the same importance. DY(t) is used for the calculus of an actuation factor on the period of each task in the following way: SDY K t DY K factor SDY t DY K factor t E K factor P P P P −=⇒−=⇒=)())((*)(*(2)where E(t) denotes the value of the error in a certain instant and it can be expressed by DY(t)-SDY, where SDY is the desirable average delay for the system. K P (proportional constant) and SDY are values defined by the designer and/or programmer of the system according to mathematical calculations or previous experimentation. A small value for K P will define a smooth actuation on the system. The factor is used by the AdaptiveService for the calculus of the effective period of each task by:)(min max min P P factor P P effective −+= (3)Besides the values of K P , SDY and DW, the mechanism for the designer/programmer to specify the actuation also includes the values of P min (minimum period) and P max (maximum period) of each task. According to the equation above, the AdaptiveService is capable of varying the period of each task (the effective period) according to the factor calculated previously, but always within the limits established by the minimum and maximum period of each task.The computation of DY(t) along the time considers that the deadline is always equal to the nominal period of each task. Along the execution, the effective period can be altered according to the adaptation factor calculated by the adaptive service. Even so, it is always used the same deadline for each task, which is equal to the respective nominal period. The adaptation is achieved by varying the task period and not the task deadline.5 Adaptive Scheduling Service Implementation on TAOThe proposed adaptation mechanism is designed for applications composed by periodic tasks. Applications can also contain aperiodic tasks and passive objects. The class diagram of one possible implementation of the mechanism can be seen in figure 1. The main elements of figure 1 are detailed in the next paragraphs.Fig. 1. Class diagram of the adaptive model implementationTaskIt represents aperiodic tasks and passive objects. It is implemented as a servant that will incarnate a CORBA object. At the moment of its creation it registers a name with the name server (NamingService) so that the active objects can find them and make method invocations.Each task is a CORBA object and it is activated in its own POA, configured with the PriorityModelPolicy as ClientPropagated, in order to respect the priority of the objects that make the invocation.PeriodicTaskIt represents the periodic tasks. This class is abstract and cannot be directly instantiated. It requires the definition of a derived class. A periodic task (PeriodicTask) is, ultimately, a consumer that enrolls in the event service (EventService) of TAO to receive time-out events in specific time intervals. This time interval is the period defined in the RT_Info of the task during the off-line phase of the scheduling.Each PeriodicTask also have an own POA and it is configured with PriorityModelPolicy as ServerDeclared (to use the priority defined in its RT_Info) and a ThreadPool with the number of static and dynamic threads defined by the programmer at the moment instances are created.AdaptiveServiceThe adaptive service (AdaptiveService) is implemented as a periodic task that needs to be described as an RT_Info in the off-line phase of scheduling. Its period must be equal to DW and can be configured in the RT_Info.The adaptive service is responsible for collecting data from the periodic tasks (the aperiodic tasks will not be considered by the service) and for posterior actuation on the system. This process can be seen in figure 2. Task PeriodicTask AdaptiveService NamingServiceEventServiceScheduleService: Inheritance : AccessIn order to use the scheduling service of TAO it is necessary to divide the work in two phases: off-line and run-time. At the off-line phase the timing properties of each task (including the adaptation service itself) are registered by using data structures called RT_Infos.At the beginning of the run-time phase each periodic task should obtain a reference for the adaptation service. It then should invoke an adaptation service method to enroll itself, also informing the values of minimum and maximum period. Each periodic task must be implemented as an instance of a subclass of PeriodicTask. That class provides mechanisms to obtain the task response time by using timers and then to invoke automatically the method mark of the adaptation service in order to inform the value measured.Fig. 2. Operation of the adaptive scheduling serviceThe adaptation service has, therefore, the function of calculating the value of DY(t) at each activation (time interval DW) and of calculating the value of the adaptation factor to be applied to the periods of all periodic tasks. The effective period calculated for each task is then used to request to the event service a new connection to each task for time-out events, but with this new value as time interval between successive activations. The response time of the adaptation service to load increases is, therefore, dependent of the response time of the event service to re-connecting the consumers for new intervals of time-out events. Situations that demand immediate response to sudden increases of load would need a fast event service when dealing with the reconnections of consumers.1.Each task connects to theevent service to receive timeout events in intervalsdefined by the periodconfigured in the RT_Info in the off-line stage of the application. 2.The adaptive service also connects to the event service to receive timeout events in accordance to therequired period. 3.Each task enrolls itself inthe adaptive serviceinforming Pmax, Pmin and also its reference.4.After all tasks being connected, the event service starts generating timeout events in intervals defined by the tasks periods (which are present in the RT Info).5. After each task execution, the response time is measured and sent to the adaptive service.6. In each activation, the adaptive service calculates the value of DY(t) within the DW window and the new factor.7. In case of the actual factor is different of the last one, the Adaptive Service requires the the event service reconnect all the tasks sending the new effective period.6 Experiences6.1 Application structureThere are many classes of distributed real-time applications that could be used to test the adaptation mechanism proposed in this work.A class of real-time applications that presents many possibilities for studying is the class of industrial automation systems. This kind of system has many different types of tasks, aperiodic, sporadic, periodic, with different levels of importance.This work sought to identify and to represent some of the types of tasks found in this type of system.•Sensor and Actuator were the elements of automation applications chosen to represent passive objects. They are derived from class Task and they have methods defined in IDL that can be invoked by active elements of the system.•We identified three possible types of periodic tasks in automation systems. They will be defined as subclasses of PeriodicTask: (i) Logger represents the tasks that are responsible for the collection of data in order to log the system status along the time, it has a period reasonably high (one second); (ii) Operator represents monitoring processes (collection and presentation of the information) that interact with a human operator; (iii) class Alarm represents monitoring processes that can identify problems in the system and they are supposed to have a period smaller than that of other tasks in order to be always updated with the last data collected in the system.The application created simulates the behavior of each class defined above and reproduces the oneway or twoway dependencies. The next section presents details on the registering of these tasks, their temporal restrictions and dependencies.6.2 Conditions of the experiencesIn order to validate the proposed adaptation mechanism, we made several experiments considering the same test scenario. After considering the application class used as test base (industrial automation), where in most cases the initial configuration is seldom altered, we assumed the same number of tasks and the same number of hardware equipments (computers, sensors, actuators, etc.) for all experiments. Experiments differ only about the details of the proposed adaptation mechanism.For all the experiments described in this section, a network formed by two machines was assumed, with the hardware and software configuration presented in table 2.Table 2. Hardware and software configuration for the experimentsMachine Hardware SoftwareNode 1 Pentium III650MHz Windows 2000 Advanced Server operating system Visual C++ 6.0 compiler128MB RAMACE version 5.2 TAO version 1.2 Node 2 AMD K7 Duron 750MHz128MB RAMWindows 2000 Advanced Server operating systemVisual C++ 6.0 compiler ACE version 5.2 TAO version 1.2Tasks were distributed among the computers in a static and arbitrary way. Table 3 shows the task allocation used in all experiments, as well as their execution times, period, dependencies registered with the TAO scheduling service and the respective priority calculated by that TAO service.Table 3. Task allocation to nodes Tasks Dependencies PriorityMachine Time Worst Exec. Time (ms)Period (ms) oneway twoway CORBA 1 Win 2k sensorA 10 - - - 2 1actuatorA 20 - - - 0 15 actuatorB 20 - - - 0 15alarm1 20 100 actuatorA actuatorB sensorA sensorB0 15Node 1 operator1 300 500 - sensorA sensorB1 2sensorB 10 - - - 2 1alarm2 20 100 actuatorA actuatorB sensorA sensorB0 15operator2 300 500 - sensorA sensorB1 2logger1 500 1000 - sensorA sensorB2 1logger2 500 1000 - sensorA sensorB2 1Adaptive Service40 1000 - - 2 1Naming Service- - - - - - Event Service- - - - - - Node 2 Schedule Service - - - - - -1 The value zero represents the higher priority and the higher value the lower priority.。
