def meanOfMeans(): initialExperiment = 25 finalExperiment = 35 suffix = "piConc" experimentFolder = "170" rootDir = os.path.join("/", "home", "rafaelbeirigo", "ql", "experiments", experimentFolder) means = [] meanFileName = "W_avg_list_mean.out" for experiment in range(initialExperiment, finalExperiment + 1): meanFilePath = os.path.join(rootDir, str(experiment), "PRQL", meanFileName) # open experiment the file that contains the "individual" mean # into a list mean = pl.loadtxt(meanFilePath) # this must be done because the function meanError.meanMultiDim() # needs each item in the mean list to be a list itself. mean = [[item] for item in mean] # append it to the list of means means.append(mean) # obtain the global mean considering the list of individual means globalMean = meanError.meanMultiDim(means) ## print globalMean # means2spreadsheet([globalMean]) pl.savetxt(os.path.join(rootDir, "W_avg_list_mean." + suffix + ".out"), globalMean, fmt="%1.6f") return globalMean
def meanOfMeans(): initialExperiment = 25 finalExperiment = 35 suffix = 'piConc' experimentFolder = '170' rootDir = os.path.join('/', 'home', 'rafaelbeirigo', 'ql', 'experiments', experimentFolder) means = [] meanFileName = 'W_avg_list_mean.out' for experiment in range(initialExperiment, finalExperiment + 1): meanFilePath = os.path.join(rootDir, str(experiment), 'PRQL', meanFileName) # open experiment the file that contains the "individual" mean # into a list mean = pl.loadtxt(meanFilePath) # this must be done because the function meanError.meanMultiDim() # needs each item in the mean list to be a list itself. mean = [[item] for item in mean] # append it to the list of means means.append(mean) # obtain the global mean considering the list of individual means globalMean = meanError.meanMultiDim(means) ## print globalMean #means2spreadsheet([globalMean]) pl.savetxt(os.path.join(rootDir, 'W_avg_list_mean.' + suffix + '.out'), globalMean, fmt='%1.6f') return globalMean
def saveOutputFiles(myQLearning, params, output): filePath = params['filePath'] myQLearning.obtainPolicy() myQLearning.printPolicy(filePath + 'policy.out') myQLearning.printQ(filePath + 'Q.out') f = open(filePath + 'parameters.out', 'w') f.write('alpha = ' + params['alpha'] + '\n') f.write('gamma = ' + params['gamma'] + '\n') f.write('epsilon = ' + params['epsilon'] + '\n') f.write('epsilonIncrement = ' + params['epsilonIncrement'] + '\n') f.write('gammaPRQL = ' + params['gammaPRQL'] + '\n') f.write('tau = ' + params['tau'] + '\n') f.write('deltaTau = ' + params['deltaTau'] + '\n') f.write('psi = ' + params['psi'] + '\n') f.write('v = ' + params['v'] + '\n') f.write('K = ' + params['K'] + '\n') f.write('H = ' + params['H'] + '\n') f.write('numberOfExecutions = ' + params['numberOfExecutions'] + '\n') f.close() #names = ['Ps'] for name in output.iterkeys(): #if name not in names: continue print name d = output[name] if len(d) > 1: dMean = meanError.meanMultiDim(d) dCfdInt = meanError.confidenceIntervalMultiDim(d, dMean) dMean = np.asarray(dMean) dCfdInt = np.asarray(dCfdInt) dMean.T dCfdInt.T np.savetxt(filePath + '/' + name + '_mean.out', dMean) np.savetxt(filePath + '/' + name + '_cfdInt.out', dCfdInt) else: d = np.asarray(d) d.T np.savetxt(filePath + '/' + name + '.out', d)