City Research Online

Abstract

In recent years supply chains have gained the attention of both academia and industry. In this thesis, a novel state-space model of a multi-node supply chain is presented, controlled via local proportional inventory-replenishment policies. The model is driven by a stochastic sequence representing customer demand. The model is analysed under stationarity conditions, guaranteed to arise if the control parameters lie in a certain range which is identified and a simple recursive scheme is further developed for updating the covariance matrix of the system in closed form, i. e., as an explicit function of the control parameters. This allows us to analyse the effect of inventory policies on the "bullwhip effect" (demand amplification) for chains with an arbitrary number of nodes. The three-node model is subsequently analysed in detail under information sharing and the optimal policy is derived, which minimises inventory fluctuations (and inventory mean) under a probabilistic constraint related to downstream demand. It is shown that this policy can never lead to demand amplification in the chain, as long as the gain parameter of the downstream node lies in the stability region. Finally, issues related to estimation schemes based on local historical data are discussed. The main results and conclusions are illustrated via numerous examples and simulations. An alternative model of the supply chain is also developed using timed Hierarchical Coloured Petri Nets (HCPN). This approach considers supply chains as event-driven systems and studies decentralised control structures by analysing the impact of various continuous inventory policies and known forecasting methods followed by supply chain participants. CPN-Tools [fCPN] are used for the design of decision-making policies and simulation results are presented to highlight the main issues arising in real systems and to provide insights for future modelling and simulation work. Finally, a detailed case study is undertaken, for the production line of the "Bridngorth Aluminium Ltd" company which produces high quality rolled aluminium lithographic strips. An efficient representation for such production processes is provided and subsequently used for an extensive analysis and performance evaluation through appropriate metrics. In particular, the work addresses the implementation of an overall model in a simulation environment, capable of integrating the various aspects of the specific production management processes.

Publication Type:	Thesis (Doctoral)
Subjects:	T Technology > TA Engineering (General). Civil engineering (General)
Departments:	School of Science & Technology Doctoral Theses School of Science & Technology > School of Science & Technology Doctoral Theses

Export

Downloads