Department of Systems Engineering & Engineering Management

The goal of this project is to provide a three-tier electronic infrastructure that allows corporations to manage their internal and external procurement/sales processes strategically. The platform will provide a user controlled interface for a real-time electronic data/document interchange, and multiple dynamic pricing models for transactions.

Analysis of Sales Promotions in Decentralized Supply Chains: Models of Monopoly and Competition

Promotions are widely used by manufacturers and retailers in varying degrees across different industries. The automotive industry in the United States is one example where production costs are largely fixed and promotions play an important role in the marketing strategies of the companies to maximize revenues. Automotive manufacturers use promotions directed to the end customers and/or retailers in their distribution channels. In this research, we develop several game theoretical models with different demand characteristics and price discriminating retailers to analyze promotions in monopoly and duopoly settings. We characterize equilibrium outcomes and determine which promotion would benefit the manufacturers under which market conditions.

Airlines and hotels tend to ally together in getting competitive edge and large market shares. Although the importance of alliance cannot be overemphasized, how to control the joint capacities and sell the combinations of services more effectively is far to be resolved. Meanwhile, since airlines seats and hotel rooms are perishing, and reserved in advance not sure to be sold, the uncertainty of many dimensions makes it even more sophisticated to manage revenues and capacities. This project considers two different but correlated problems. One is the revenue management on the combinations of products, assuming the capacities and prices are given. These combinations can be thought of as a seat in several connecting flights, or a seat in several flights and a hotel room in the destination. Selling them not only affects their capacities alone but also the capacities of other combinations. The second interesting problem is the interface between the customers and product combinations. It aims at exploiting adequate data of customer preferences to sell the right combinations at the right times to the right customers. This project is tackling the fundamentals of current revenue management research and development.

Coordinated Decisions of Manufacturer/ Distributor in a Fresh Product Supply Chain Involving Long Distance Transportation

We consider a supply chain where a manufacturer produces a variety of fresh products to supply a distributor in a distant export market. The manufacturer faces the risk that a fresh product may decay during the process of long distance transportation, in particular in the presence of uncertain events (such as bad weather, airport delays, etc.). The distributor faces the risk that the demand for a product is uncertain and any unsold fresh product may lose its value after the sales period. While the profit potential in supplying the products to the export market is substantial, a great challenge for both parties is how to minimize the losses involved. Because time is a crucial element for fresh products, proper decisions regarding the timing to produce, deliver, and sell, become particularly critical in these situations. Main topics to be investigated include modeling to capture the prominent characteristics and concerns in different scenarios, derivation and analysis of optimal policies, and design and analysis of information and profit sharing schemes.

It is well known that commodities, such as oil, natural gas, electricity, metals, are traded frequently in the exchanges in the standard futures contracts or under-counters in the mutually agreed forwards. They cannot be regarded as simple financial assets since they will be consumed, stored and transported around the world. This project copes with pricing the transportation of commodities. For instance, an oil tanker can be used for storing and transferring oil. It can be used by an oil refinery to hedge against the risk of oil price fall-down. It can be also used by a speculator to make profit by arbitraging the price spreads between geographical places and time points. Thus, the capacity of an oil tanker and pricing on it is not a matter of simple transportation. Its pricing is driven by market forces and it has to move around in order to sufficiently utilize the capacity. The goal of this project is how to price the transportation commodities properly, and how to make commodity transportation companies more financially stable, given random price spread and demand.

Dynamic Pricing Problems for Seasonal Products with Inventory-Record Inaccuracy: a Revenue Management Approach to Exploring the Value of RFID

Contrary to popular belief, modern retailers that are equipped with advanced information technology systems do not actually have an accurate record of the products that they have in stores and warehouses, and recent empirical studies have shown that unaccounted inventory discrepancies have a serious effect on operating costs and revenue. As a mobile technology, RFID (radio-frequency identification) can bring high visibility to supply chains and give an accurate account of inventory. Our objective is to develop methods to help solidify the quantification of the value of RFID. The main subject of the proposed project is seasonal or fashionable products that are sold in an environment in which inventory information is inaccurate. Specifically, this project will take a step toward the economic justification of RFID by assessing costs and revenue with the current inaccurate inventory information and costs and revenue when inaccurate information is reduced or eliminated through RFID adoption. Therefore, this project will focus on the value of RFID in the better management of inventory systems with inventory inaccuracies.

F. Chen and Y.Y. Feng

The bundling of service packages and/or accessories with the sale of commodities has become an important source of profit in many industries. Different packages can be bundled together to form different offerings that have varying attractions for different customer segments. Hence, what packages or subsets of packages (bundles) to offer over time will affect a firm's sales and revenue. This should be reflected in the management of inventory of commodities, which is the purpose of this project. We consider two models. In the first model, a single product is sold with value-added service packages, and its inventory can be replenished at any time. The objective is to optimally determine the stock replenishment and package offering. In the second model, we consider seasonal products which can be sold as bundles or separately, and there is only one ordering opportunity. The objective is twofold: to determine the initial ordering quantity for each product and to decide the bundle offering at any point in time (before the season ends).

Inventory Management of Hybrid Remanufacturing/Manufacturing Systems with Multiple Types of Return Products

Nowadays, product return flow is becoming a significant concern of many manufacturers. In the flow, there usually are return products bearing different conditions, which most likely will incur different remanufacturing and sometimes storage costs. The serviceable products used to fulfill customer demand can be either manufactured from raw material or remanufactured from different types of return products. We develop and analyze remanufacturing/production planning and inventory management for periodic-review systems with stochastic demand and multi-type return products. The objective is to characterize the optimal inventory policy that minimizes the system cost. The computational procedure of the optimal control parameters and some simple heuristics are also studied.

Along this line, we also plan to carry out the following research projects. One is to consider the systems with different remanufacturing and production leadtimes. The optimal policy will become very complicated but the question is whether some simple effective heuristics could be developed. Another potential research direction is to develop models that the remanufacturing process is divided into different phases. We aim to show the optimal inventory policies. At last, if manufacturers can actively collect the returns with proper strategies, how to manage such hybrid inventory systems.

J. Leung and C. H. Cheng

The goal of this project is to develop a modular delivery-planning system that supports and exploits technologies for Mobile-commerce, and is adaptable to handle the traffic congestion and population density of Hong Kong. The target system will provide location-based information, on-line query/ordering and delivery-tracking for customers. It will also support delivery planning (order assignment and routing) and dynamic re-routing using time-based information. The system will be modular and will contain bilingual digital maps and traffic databases of Hong Kong. The key research questions to be addressed in this project are: database design to accommodate and exploit dynamic updating and global-positioning, and algorithmic development of routing and scheduling methods (optimal or heuristic) for dynamically-updated databases and changing routing/scheduling specifications.

Firms are exposed to a wide variety of risks, such as uncertainties about the supply of key inputs and their prices, and exchange rate fluctuations. Corporate risk management programs aim to systematically manage such risk exposure to increase firm value. This project is concerned with market risks that can be partially hedged through financial instruments in general, and weather risk - the uncertainty in cash flow and earnings that is caused by weather volatility, and price risks in particular. We will study the interactions of the operational decisions and risk hedging strategies of affected firms.

In the global competition of supply-chains vs. supply-chains, swiftness in each value-added process is essential. Many time-critical steps involve rapid "servicing" of transportation equipment at consolidation hubs. This project focuses on crew-scheduling for such servicing to ensure rapid turnaround-times in the supply-chain. This job-assignment/scheduling problem is highly complex because of: the extremely tight time-windows for servicing, and large variety of equipment types, making matching appropriately-skilled crews to jobs difficult. The possibility of new operational modes -- such as sharing of jobs by multiple crews to reduce service times - adds further complications. We hope to extend our understanding of workforce flexibility in rostering/scheduling models. At the operational level, we will develop effective job-assignment and scheduling methods that allows job-sharing. At the strategic level, we explore how cross-training and broader skill-sets of the crews impacts optimal schedule costs. In actual operations, unanticipated delays occur, so disruption recovery and redeployment of crews are necessary. This project will also investigate methods for stochastic planning and disruption recovery. Results from this project may be relevant to many logistics businesses whose operations are time-critical. We hope our findings will contribute to understanding of the value of resource flexibility.

Manufacturing Logistics with Application to Electronic Equipment Assembly and Distribution

Manufacturing logistics process re-engineering is the search for and implementation of radical change in manufacturing logistics processes to achieve breakthrough results. In this project, the researcher plans to analyze and develop a supply chain model for the assembly and distribution of electronic equipment and devices, by re-engineering manufacturing logistics processes. Several features distinguish our problems from the traditional mode: unprecedented number and variety of products, short product life cycles and forecasts of volatile demands are quite inaccurate. Furthermore, the manufacturing efficiency strongly depends on suppliers of key components. How to reduce the lead-time and finished-product inventory without downgrading the customer service is important. This research addresses three issues. First, we will build a model that captures the basic blocks of the supply chain including raw material and components supply, product assembly, and finished product distribution. The model serves as a tool to validate and/or compare strategies and alternatives. Second, we will analyze the impact of process, especially re-engineering manufacturing processes, on lead-time and costs. The quantitative assessment serves as a guideline to predict the necessity of process re-engineering. Third, we will develop a procedure to assign key parameters for the supply chain in an optimal or near optimal fashion. The research will lead to effective methods that provide strategic and operational benefits to regional industries.

Multi-Echelon Stochastic Inventory Systems: Optimality, Optimization, Bounds and Heuristics

Multi-echelon inventory system is one of the most important research streams in supply chain management. In this research, we analyze several multi-echelon inventory systems including system with batch ordering and nested replenishment schedule, system with regular and expedited shipping, etc. We characterize the optimal inventory policies and provide optimization algorithms for computing the optimal policies. Bounds for cost functions and the optimal policies are developed, which leads to some simple near-optimal heuristics. The analytical results improve the transparency of the system and enable us to investigate the system design issues.

Pricing, Production and Delivery Decisions, and Cooperative Strategies in a Supply Chain with Products of Time-Varying Values

Many industries face the problem of manufacturing and selling products of time-varying values. Due to the time-varying nature of product values, determining the proper decisions and strategies regarding the best timing to offer new sales price, to place order, and to produce and deliver, is a great challenge for the manufacturer as well as the retailers involved in the supply chain. In this project we examine a supply chain with one manufacturer and multiple retailers, where the manufacturer wishes to determine a proper pricing mechanism and the corresponding production/delivery decisions, while the retailers wishes to make use of the pricing mechanism offered by the manufacturer, through possible grouping with each other to reach the needed purchase quantities for price drops. Cooperation and competition among the retailers, and between the manufacturer and the retailers, will be considered.

Studies of Combined Segmental-dynamic Pricing and Production/Replenishment Decisions

F.Y. Chen and Y.Y. Feng

Using different prices among customer segments and over time has long been recognized as a powerful revenue management tool. One issue that arouses the interest of researchers and practitioners is the application of revenue management to the production/distribution environment, where the resources may be non-perishable and renewable. We plan to extend the idea of differentiating prices according to customer segment and time-period in revenue management to the production environment. The focus of the research is on a class of problems that exist in a wide range of applications, and has a certain structure that entails tractable solutions.

Modeling, analysis and optimal design of large-scale supply networks. Coordinating quality control schemes and production/replenishment decisions in multi-stage systems. The objective is to develop methods and technologies that support decision making in supply chain management.

This project, carried out in cooperation with Statoil, one of the world's largest oil companies, focuses on delivering goods and people to off-shore oil and gas installations in the Norwegian Sea. There are rough waters, where uncertain weather (in particular wave heights) can prevent deliveries during rather long periods, especially in winter. This project relates inventory theory to vehicle routing under extremely high shortage costs.

C.H. Cheng and J. Leung

A fundamental problem in crossdocking operations is concerned with the arrangement of strip doors (where incoming freight is unloaded from a trailer) and stack doors (where outgoing freight is loaded onto a trailer), and the assignment of destinations to stack doors. This is referred to as the layout problem in a crossdock. To the best of our knowledge, most existing work only deals with the static problem in which the layout remains the same for the planning horizon. The static treatment provides good analytical implications. However, if facilities operate under a very dynamic environment, layout rearrangement may be required during the planning horizon to maintain layout effectiveness. In this project, we attempt to study the dynamics of layout in a crossdock. We address several research issues of crossdocking operations that have not received sufficient attentions:

• Dynamic aspects with re-layout cost,

• Impacts of different material handling equipment,

• Interactions of layout, freight composition, and congestion.

This problem is particularly important for Hong Kong because of the high cost of land. Warehouse spaces and loading areas are severely limited in most distribution centre operations. As part of our work, we will develop models and solutions applicable to the Hong Kong situation.

J. Leung and C.H. Cheng

Transportation logistics in Hong Kong is very complex. The highly unpredictable transit times due to traffic congestion and/or accidents, the tight time-windows and high reliability demanded by the customers make delivery planning a nightmare for the hundreds of logistics service-providers in Hong Kong. In this project, we will develop a Traffic Inference Engine for transit-time estimation for all road segments in a network, based on information on the durations of trips between various origins and destinations points. The core of the engine is a mathematical-programming model which can be adapted to various forms and frequencies of data input, and for which computations for estimation-updating can be done rapidly. The technology is also scalable, adaptable for use over the entire road network in Hong Kong or applied to a more focussed region of interest to a particular organisation.

C.H. Cheng, W.M. Cheung, and J. Yeung

An ERP system is a new management technology that advocates an integrated approach to conduct business. While organizations are hoping to apply this technology to improve overall performance, they must understand what it takes for their employees to use it. Although the use of ERP systems may not be voluntary, the understanding of system adoption from the user's perspective is useful in helping the organizations prepare their employees to face new challenges and leam how to make good use of the technology. To analyze factors affecting the ERP system usage, we proposed a conceptual model derived from the Triandis framework. The use of the Triandis framework is based on the previous research that documents the importance of social factors on the adoption of a technology. An empirical study was conducted in Hong Kong to understand the adoption process. Based on our findings, we also propose managerial implications in connection with promoting the usage.