No description available.

No description available.

No description available.

At the time of a customer order, the e-tailer assigns the order to one or more of its order fulfillment centers, and/or to drop shippers, so as to minimize procurement and transportation costs, based on the available current information. However this assignment is necessarily myopic as it cannot account for all future events, such as subsequent customer orders or inventory replenishments. We examine the potential benefits from periodically re-evaluating these real-time order-assignment decisions. We construct near-optimal heuristics for the re-assignment for a large set of customer orders with the objective to minimize the total number of shipments. We investigate how best to implement these heuristics for a rolling horizon, and discuss the effect of demand correlation, customer order size, and the number of customer orders on the nature of the heuristics. Finally, we present potential saving opportunities by testing the heuristics on sets of order data from a major e-tailer. / Singapore-MIT Alliance (SMA)

Logistics and supply chain problems arising in the context of intermodal transportation and less-than-truckload (LTL) network design typically require commodities to be consolidated and shipped via the most economical route to their destinations. Traditionally, these problems have been modelled using network design or hub-and- spoke approaches. In a network design problem, one is given the network and flow requirements between the origin and destination pairs (commodities), and the objective is to route the flows over the network so as to minimize the sum of the fixed charge incurred in using arcs and routing costs. However, there are possible benefits, due to economies-of-scale in transportation, that are not addressed in standard network design models. On the other hand, hub location problems are motivated by potential economies-of-scale in transportation costs when loads are consolidated and shipped together over a completely connected hub network. However, in a hub location problem, the assignment of a node to a hub is independent of the commodities originating at, or destined to, this node. Such an indiscriminate assignment may not be suitable for all commodities originating at a particular node because of their different destinations. Problems arising in the area of LTL transportation, intermodal transportation and package routing generally have characteristics such as economies- of-scale in transportation costs in addition to the requirement of commodity-based routing. Obviously, the existing network design and hub location-based models are not directly suitable for these applications. In this dissertation, we investigate the development of models and solution algorithms for problems in the areas of LTL and intermodal transportation as well as in the freight forwarders industry. We develop models and solution methods to address strategic, tactical and operational level decision issues and show computational results. This research provides new insights into these application areas and new solution methods therein. The solution algorithms developed here also contribute to the general area of discrete optimization, particularly for problems with similar characteristics.

Since the introduction of the Internet and the realisation of its potential companies have either transformed their operation or are in the process of doing so. It has been observed, that developments in I.T., telecommunications and the Internet have boosted the number of enterprises engaging into e-commerce, e-business and virtual enterprising. These trends are accompanied by re-shaping, transformation and changes in an enterprise's boundaries. The thesis gives an account of the research into the area of dynamic enterprise modelling and provides a modelling methodology that allows different roles and business models to be tested and evaluated without the risk associated with committing to a change.

Australia has a widespread and significant incidence of land and water contamination, which can lead to economic, trade, ecosystem and human health impacts. Over the past 20 years the problem has been growing and there is also a growing realization of the extent of the problem. The installation of monitoring network is pivotal for understanding the groundwater hydraulics and subsurface contamination. At the same time the process is expensive. The systematic study of the subsurface system with the available scanty data regarding the groundwater flow and the subsurface contamination can help us to arrive at the optimum monitoring network design for effective site characterization.

Feedforward artificial neural networks (FANNs), which have been successfully applied to various image processing tasks, are particularly suitable for image subsampling due to their high processing speed. However, the performance of FANNs in image subsampling, which depends on both the FANN topology and the FANN training algorithm, has not been acceptable so far. High performance image subsampling is important in many systems, such as subband decomposition systems, and scalable image and video processing systems. This thesis addresses the design of FANNs with application to image subsampling. More specifically, we focus on both the topological design of FANNs and the training algorithm, so that efficient FANN structures, yielding good performance in image subsampling, are obtained. That is, we aim at obtaining compact FANNs that yield good subsampled versions of the original images, such that if reconstructed, they are as close as possible to the original images. Moreover, we aim at obtaining better performance-speed tradeoffs than those of the traditional lowpass filtering and subsampling methods. First, we propose a design method for FANNs, which leads to compact tridiagonally symmetrical feedforward neural networks (TS—FANNs). Next, in order to address the problem of artifacts that generally appear in the reconstructed images after FANN-based subsampling, we propose a training method for FANNs. When applied to first-order (FOS) and multi-stage first-order (MFOS) image subsampling, the FANNs trained using our method outperform the traditional lowpass filtering and subsampling (LPFS) method, without requiring pre- or post-processing stages. Motivated by our observation that the computational demands of the MFOS process increase approximately linearly with the image size, we then combine the proposed methods and evaluate the performance-complexity tradeoffs of the resulting TS-FANNs in FOS and MFOS. We show that our TS-FANNs-based subsampling has important advantages over subsampling methods based on fully connected FANNs (FC—FANNs) and LPFS, such as significantly reduced computational demands, and the same, or better, quality of the resulting images. The main contributions of this thesis consist of a method for FANN design with tridiagonal symmetry constraints, a training algorithm for FANNs applied to image subsampling, the design and evaluation of the performance-speed tradeoffs of FC—FANNs in image subsampling, and the design and evaluation of the performancespeed tradeoffs of TS—FANNs in image subsampling. The FANN performance in image subsampling is evaluated objectively (using the peak signal-to-noise ratios), subjectively (by visual examination of the subsampled and of the reconstructed images), and in the context of a video coding application. The speed and memory demands of the designed FANN structures are evaluated in terms of the subsampling time and the number of FANN parameters, respectively. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

No description available.

On April 27, 1802, I gave a shout of joy ... It was seven years ago I proposed to myself a problem which I have not been able to solve directly, but for which I had found by chance a solution, and I knew that it was correct, without being able to prove it. The matter often returned to my mind and I had sought twenty times unsuccessfully for this solution. For some days I had carried the idea about with me continually. At last, I do not know how, I found it, together with a large number of curious and new considerations concerning the theory of probability. Andre Marie Ampere. Facility location problems (or plant location problems) are general models that can be used when a set of clients has to be served by facilities. More precisely, we are given a set of potential facility locations and a set of clients. The optimization problem is to select a subset of the locations at which to place facilities and then to assign clients to theses facilities so as to minimize total cost. Most formulations considered in this thesis can be viewed as general models that can be applied to a wide range of context and practical situations. However, as this research has been partly initiated by the interest of the author in telecommunication network design we will introduce these models by considering problems in this particular area. In the context of telecommunication network design an application of discrete location theory is the optimization of access networks with concentrators. Typically, we have a number of terminal points that must be connected to a service point. An obvious solution is to use a dedicated link for each terminal (star network). However, it is clear that this solution can be very expensive when the number of terminals is large and when they are far from the service point. Access networks are often constructed by inserting concentrators between the terminals and the service point. Many terminals are connected to a facility which in turn is connected by a single link to the service point. The objective is to build a network that will provide the service at minimum cost. If no extra constraints are involved the mmimum cost network problem can be expressed as an uncapacitated facility location problem (UFL). If the number of terminals that can be connected to a concentrator is limited we obtain a so-called capacitated facility location problem (CFL). CFL can be extended to consider various types of concentrators with various capacities. This problem is the multi-capacitated facility location problem (MCFL). MCFL is a straightforward model for low speed packet switched data networks typical among which are networks connecting sellingpoint terminals to a database. For other networks, the problem may involve various traffic constraints. In chapter 1 we present those problems and compare solutions obtained by Lagrangian relaxation and simulated annealing algorithms. The architecture mentioned above can be extended with more than one hierarchical level of concentrator. Unfortunately, we pay for this cost saving through a decrease of reliability. Therefore, the number of levels is often limited to one or two. In chapter 2 we study an extension of UFL and CFL to two levels of concentrators. Obviously, the structure of a network changes according to the way requirements vary with time. In order to plan investments and to develop strategies, the evolution of a network has to be determined for several years ahead (typically four or five years). In this case the main questions to answer are: Where and when to establish concentrators and of what size? In chapter 3 we study this problem for the dynamic version of UFL. Now, with the network optimization problem, there naturally arises the problem of allocating the total minimum cost among customers fairly. Namely, we would like to allocate the cost in such a way that no subgroup of users would have incentive to withdraw and build their own network. The standard way to approach such a problem is by the means of cooperative game theory. In chapter 4 we study the core of location games derived from UFL and CFL, and in chapter 5 we propose methods to compute the nucleolus of these games.