Dissertations/Theses - Department of Industrial and Production Engineering

Browse

Now showing 1 - 3 of 3

Integrated economic design of quality control and maintenance management using CUSUM chart with VSIFT sampling policy
(Department of Industrial and Production Engineering, 2017-10-16) Saha, Rajesh; Azeem, Dr. Abdullahil
Due to the close interrelation between statistical process control and maintenance management policy and the necessity of these two key tools in running a smooth production system, this paper presents an integrated economic model for joint optimization of quality control parameters and preventive maintenance policy with cumulative sum (CUSUM) control chart. In this model CUSUM chart is used to monitor both process mean and variance and joint average run length (ARL) is determined by combining mean and variance through absorbing Markov chain approach. Here the CUSUM chart is designed using variable sampling interval fixed time (VSIFT) sampling policy. In order to determine the in control and out of control chart for both mean and variance, Taguchi quadratic loss function and modified linear loss function are used respectively in this model. In this proposed model two types of maintenance policy i.e. imperfect preventive maintenance and minimal corrective maintenance have been considered. The proposed model determines the optimum values of eight test parameters (the sample size (n), the fixed sampling interval (h), the number of subintervals between two consecutive sampling times (ȵ), the control limit coefficient for CUSUM mean chart (k), the warning limit coefficient for CUSUM mean chart (w), the time interval of preventive maintenance (tpm), the control limit coefficient for CUSUM variance chart (k1) and the warning limit coefficient for CUSUM variance chart (w1)) so that expected total cost per unit time is minimized. A numerical example is presented to demonstrate the effectiveness of the test model in cost minimization. Nelder-Mead downhill simplex method and Genetic algorithm approaches are applied to search for the optimal values of the eight test parameters for the economic statistical design of VSIFT CUSUM charts. A sensitivity analysis has also been performed to observe the effect of different process parameters on total cost.
Resilience-based network design optimization under epistemic uncertainty
(2019-07-29) Sharmin, Tavila; Zaman, Dr. AKM Kais Bin
Design optimization of infrastructure networks is a crucial task due to its huge impacts in socio-economic sectors and consideration of resilience in the design phase of a network ensures a higher level of performance even after the occurrence of any disruptive event. Due to the lack of knowledge regarding the effects of future disruptive events and the exact values of the resilience parameters, uncertainty consideration is essential for this type of optimization formulation. In this research, resilience-based network design optimization models are formulated with a view to minimizing design cost, maximizing the level of performance in the post-disrupted state, and maximizing resilience. The models are formulated for both deterministic and stochastic cases. The stochastic model, formulated in this study, is able to deal with epistemic uncertainty arising from interval data of the resilience parameters. The solution methodology generates the optimal network topology and the design capacities of each link present in the optimal solution. Finally, the formulations are solved to generate the optimal network that can satisfactorily perform even after multiple disruptive events.
Robust and reliability-based design optimization under epistemic uncertainty
(Department of Industrial and Production Engineering, 2015-01) Dey, Prithbey Raj; Kais Bin Zaman, Dr. A K M
This thesis proposes formulations and algorithms for robust design optimization with uncertainty representation and propagation considering both aleatory (e.g. produced due to natural variability) and epistemic (e.g. variability due to lack of information or imprecise information) uncertainty arising from interval data. Multiple interval data are treated for uncertainty representation including both overlapping and non-overlapping in characteristics. A general likelihood-based approach for uncertainty representation has been proposed in this research. Uncertainty analysis through the likelihood approach is capable of estimating the uncertainty for different distribution types and parameters. The proposed likelihood-based representation of epistemic uncertainty has been used in the framework for robustness-based design optimization to achieve computational efficiency. A methodology is also outlined for solving reliability-based design optimization (RBDO) under epistemic uncertainty using the proposed likelihood-based uncertainty representation. The proposed robust design optimization methodology is illustrated with two numerical examples including a general mathematical problem and a real engineering problem.