Sharif Digital Repository

This effort attempts to study the change point problem in the area of non-linear profiles. Two methods for estimating the time of a single step change is proposed. In the first method a model consisting of two networks which is based on artificial neural networks is proposed. These networks are different only in their training data. One network is trained for ascending segments of the profile and the other is trained for descending segments of the profile. In the second method the maximum likelihood estimator (MLE) of the single step change is analyzed. Due to the complexity of estimating the parameters of the non-linear model by MLE, this estimator is based on the difference between the... 

This effort attempts to study the multiple change-point problem in the area of non-linear profiles. Two methods for estimating the times of multiple change-points is proposed. In the first method, a model consisting of two networks, which is based on artificial neural networks, is proposed. These networks are distinctive only in their training data. One network is trained for ascending segment of the profile and the other is trained for descending segments of the profile. In the second method, Bayesian approach is proposed for estimating multiple change-point. While using Bayesian approach the parameters of the Non-linear model must be estimated. However, this issue is complicated or... 

The aim of this research is to study performance of Rao Score Test control chart in phase II monitoring of generalized linear profiles for drift and multiple changes which can be isotonic or antitonic. Moreover, the performance of the method is compared with two common methods in the generalized linear profile literature: Hotelling T2 and multivariate exponential weighted moving average. Afterward multivariate cumulative sum chart is proposed to be used in monitoring antitonic multiple change in the parameter of Poisson profiles. Finally, a real world example is presented in which Rao Score Test method is applied to real data and the performance of this method is compared with other methods.... 

Control charts are one of the strongest optimization tools. Control charts issue warning due to out of control processby recorded data. As soon as charts warn, attempt start to find the changes reason.Finding out the issue on time save a lot of time and cost. By determining the time of this change, finding out the problem reason get faster. The real time of process change called change point. There exist many papers in change point field finding real time of change in literature. Each one considers the problem with specific assumptions. These assumptions consist of distribution function, change type, parameters and solution procedure. One kind of existing papers are about to determine normal... 

Control charts are one of the most important statistical process control tools used in monitoring processes and improving the quality by decreasing the variability of processes. In spite of various applications for multi-attribute control charts in industries and service sectors, only a few research efforts have been performed in developing this type of control charts. The developed multivariate control charts are all based on the assumption that the quality characteristics follow a multivariate Normal distribution while, in many applications the correlated quality characteristics that have to be monitored simultaneously are of attribute type and follow distributions such as multivariate... 

Control charts are one of the most important tools in statistical process control to detect assignable causes. The goal of a control chart is to detect an out-of-control state quickly so that process engineers can initiate their search for the special cause sooner. Once the special cause has been identified, the appropriate action can then be taken to improve the process.A new method to estimate the change point of Poisson rate parameter when the step change occurs is proposed in this thesis. That is, the rate parameter is assumed to suddenly shift from its in-control value to an out-of control value at a single unknown point in the process.To do this, a belief that the process is in-control... 

In this research, we estimate the change point in multistage processes using maximum likelihood estimation approach. We first model a multistage process with one quality characteristic in each stage, with both AR(1) and ARMA(1,1) time series model and then a maximum likelihood estimator for linear trend change point is developed. Also, a multivariate multistage process is modeled with VAR(1) time series model and the step change point is estimated using maximum likelihood estimator for multivariate multistage process  

Control charts are considered as one of the most important tools of statistical process control in detection of assignable causes of variation in the processes. One of the main criticisms of these charts is their inability in discovering the out-of-control state in real time. To eliminate the main sources of error, indicating the actual time of deviation in processes which is called change point is very important. Diagnosing of real time of changes limits the range of search for the causes of deviations and maximizes the chance of finding the main sources of deviation resulting in time saving and reducing expenses. There are different types of change points. One of change point types which... 

Control charts are the most important tools of statistical quality control. The problem that exists is that control charts do not show the real time the shift in a process started. The real time a change occurs in a process is called the change point. In this research, the monotonic change point in a multivariate normal process is estimated using the maximum likelihood estimation approach, where the process is monitored by a multivariate exponentially weighted moving average scheme. Mote Carlo simulation studies are performed to evaluate the performance of the proposed approach  

Knowing the time of change would narrow the search to find and identify the variables disturbing a process. Having this information, an appropriate corrective action could be implemented and valuable time could be saved. Multistage processes that are often observed in current manufacturing processes must be monitored to assure quality products. The change-point detection of such processes has not been proposes investigated yet. Thus, this dissertation proposes maximum likelihood step-change estimators of two kinds of these processes. First, a multistage process with variable quality characteristics is considered and formulated by the first-order auto-regressive model. For the location... 

Control charts are the most popular Statistical Process Control tools used to monitor process changes. However, they are not capable of identifying the real time of a process change, which is essential for diagnosing assignable causes of the change. Therefore, a number of methods of change-point estimation have been developed. In the literature, relatively little study has been done on multiple changes. In this research, a new method based on Maximum Likelihood Estimator (MLE) is introduced to identify linear drifts and step changes in multivariate multiple linear profiles. Due to the massive increase in the amount and time of the calculations along with the growth of the number of the... 

There are many cases in industrial and non-industrial sections where the quality characteristics are in the form of profiles. A profile is the functional relationship between a response variable and one or more predictor variables used to describe the quality of a process. Profile monitoring is the implementation of statistical process control techniques for this purpose. According to the type of relationship between response variable and predictor variables, profiles are classified into many categories such as: simple linear profiles, multiple linear profiles, nonlinear profiles and generalized linear profiles. Most of the research efforts in the area of profile monitoring have been...