Michael Kuby

The hub network design problem with stopovers and feeders: The case of Federal Express

This paper introduces a network planning problem called the hub network design problem with stopovers and feeders. Most, if not all, of the recent logistics research on hub-and-spokes networks has assumed that all nodes in the network are connected by direct flights to the hub. However, in the network used by the Federal Express Co., most flights to and from the hub make one or more stopovers, and many smaller cities are served by feeder aircraft which connect to other nonhub cities. This paper explores the tradeoffs and savings involved with stopovers and feeders, and develops a mixed-integer program to design the least-cost single-hub air network assuming that the hub location is already determined. The model is then applied to the western U.S. portion of the Federal Express package collection system. Comparing the optimal results to the pure hub-and-spokes network proves that substantial improvements in cost, miles flown, load factor and number of aircraft can be achieved by using stopovers and feeders in a hub network, and that it is unrealistic to assume a network with only direct flights. Comparison with the actual Federal Express network shows many similarities (which suggest that the model is capturing the important design criteria), and several differences (which indicate the models potential for improving efficiency). The usefulness of the model for a companys comprehensive network planning and for hub location modeling is discussed.

Heuristic algorithms for siting alternative-fuel stations using the Flow-Refueling Location Model

This paper presents three heuristic algorithms that solve for the optimal locations for refueling stations for alternative-fuels, such as hydrogen, ethanol, biodiesel, natural gas, or electricity. The Flow-Refueling Location Model (FRLM) locates refueling stations to maximize the flow that can be refueled with a given number of facilities. The FRLM uses path-based demands, and because of the limitations imposed by the driving range of vehicles, longer paths require combinations of more than one station to refuel round-trip travel. A mixed-integer linear programming (MILP) version of the model has been formulated and published and could be used to obtain an optimal solution. However, because of the need for combinations of stations to satisfy demands, a realistic problem with a moderate size network and a reasonable number of candidate sites would be impractical to generate and solve with MILP methods. In this research, heuristic algorithms--specifically the greedy-adding, greedy-adding with substitution and genetic algorithm--are developed and applied to solve the FRLM problem. These algorithms are shown to be effective and efficient in solving complex FRLM problems. For case study purposes, the heuristic algorithms are applied to locate hydrogen-refueling stations in the state of Florida.

An arc cover–path-cover formulation and strategic analysis of alternative-fuel station locations

In this study, we present a new formulation of the generalized flow-refueling location model that takes vehicle range and trips between origin–destination pairs into account. The new formulation, based on covering the arcs that comprise each path, is more computationally efficient than previous formulations or heuristics. Next, we use the new formulation to provide managerial insights for some key concerns of the industry, such as: whether infrastructure deployment should focus on locating clusters of facilities serving independent regions or connecting these regions by network of facilities; what is the impact of uncertainty in the origin–destination demand forecast; whether station locations will remain optimal as higher-range vehicles are introduced; and whether infrastructure developers should be willing to pay more for stations at higher-cost intersections. Experiments with real and random data sets are encouraging for the industry, as optimal locations tend to be robust under various conditions.

Technological Change and the Concentration of the U.S. General Cargo Port System: 1970–88

AbstractThe diffusion of containerization has changed not only how general cargo is handled, but where. Using the Gini coefficient, we show that general cargo port traffic has become more concentrated from 1970 to 1988 because of four technological changes: containerization, larger ships, larger trains, and computerization of freight tracking and billing. These four technological changes have spawned four kinds of intermodal services: microbridge, minibridge, landbridge, and round-the-world. We reconcile this concentration trend with Hayuths (1988) seemingly contradictory finding that containerized cargo, which makes up most of general cargo, became less concentrated throughout the U.S. port system from 1970 to 1985. Anticipated future technological innovations are expected to continue the concentration trend. Our results fit well into Slacks (1990) proposed addition of a seventh stage (dropping of redundant nodes) to the Taaffe, Morrill, and Gould (1963) model of network development. In a methodologica...

An efficient formulation of the flow refueling location model for alternative-fuel stations

The Flow-Refueling Location Model (FRLM) locates a given number of refueling stations on a network to maximize the traffic flow among origin–destination pairs that can be refueled given the driving range of alternative-fuel vehicles. Traditionally, the FRLM has been formulated using a two-stage approach: the first stage generates combinations of locations capable of serving the round trip on each route, and then a mixed-integer programming approach is used to locate p facilities to maximize the flow refueled given the feasible combinations created in the first stage. Unfortunately, generating these combinations can be computationally burdensome and heuristics may be necessary to solve large-scale networks. This article presents a radically different mixed-binary-integer programming formulation that does not require pre-generation of feasible station combinations. Using several networks of different sizes, it is shown that the proposed model solves the FRLM to optimality as fast as or faster than currently utilized greedy and genetic heuristic algorithms. The ability to solve real-world problems in reasonable time using commercial math programming software offers flexibility for infrastructure providers to customize the FRLM to their particular fuel type and business model, which is demonstrated in the formulation of several FRLM extensions.

Generating candidate networks for optimization: The CO2 capture and storage optimization problem

We develop a new framework for spatially optimizing infrastructure for CO2 capture and storage (CCS). CCS is a complex and challenging problem: domestically deploying CCS at a meaningful scale will require linking hundreds of coal-fired power plants with CO2 sequestration reservoirs through a dedicated and extensive (many tens-of-thousands of miles) CO2 pipeline network. We introduce a unique method for generating a candidate network from scratch, from which the optimization model selects the optimal set of arcs to form the pipeline network. This new generation method can be applied to any network optimization problem including transmission line, roads, and telecommunication applications. We demonstrate the model and candidate network methodology using a real example of capturing CO2 from coal-fired power plants in the US Midwest and storing the CO2 in depleted oil and gas fields. Results illustrate the critical need to balance CCS investments with generating a candidate network of arcs.

PROACTIVE OPTIMIZATION OF TOXIC WASTE TRANSPORTATION, LOCATION, AND TECHNOLOGY.

Abstract Many models of real world problems, such as the toxic waste transportation and location problem, produce solutions that “make the best of a bad situation”. Yet in many cases, giving the model better choices with which to work could produce far superior results. We introduce a framework for proactive optimization, defined as identification of the structural parameters within an OR problem that cause optimal solutions to be less than satisfactory, followed by an exogenous search for better options to add to the model. The proactive methodology is illustrated by a multiobjective, mixed-integer, location-allocation model with technology choice variables. A new technology, solar-driven waste detoxification, is compared with toxic waste incineration on three traditionally conflicting criteria: cost, a new risk measure (μg/m3 person hrs), and a new disequity measure (MinMaxSum kg∗km). The solar process is found to improve all three objectives considerably. Sensitivity analysis indicates the robustness of the results in terms of cost, risk, and sunlight availability.

A network transformation heuristic approach for the deviation flow refueling location model

In the early stages of development, alternative-fuel vehicles will tend to have shorter driving ranges than conventional vehicles, and the availability of stations will be limited. Given these conditions, it is important to consider the willingness of drivers to deviate to some extent from their shortest paths in order to refuel their vehicles and complete their trips. Previously, we proposed the deviation-flow refueling location model (DFRLM) for locating a given number of refueling facilities to maximize the total alternative-fuel vehicle flows that can be refueled by drivers traveling on or deviating from their shortest paths. On a real-world problem, however, the large number of possible deviations from each path and of combinations of facilities that can cover each path would make it extremely difficult to generate and solve the mixed-integer formulation. This paper develops heuristic algorithms for the DFRLM that overcome this difficulty through network transformation. Specifically, a greedy heuristic constructs and edits an artificial feasible network in which each node represents a station, origin, or destination, and each arc represents a feasible path between two nodes given the assumed driving range of vehicles. At each step of the greedy and greedy-substitution algorithms, the feasible network is edited and a shortest path algorithm is run, which determines whether each origin-destination round trip can be completed. This method allows any possible detour to be taken (up to some user-defined maximum) while also ensuring that drivers take the smallest possible detour. Computational experiments on a simple network and a real-world network for Florida show the heuristics to be efficient in solving the problems. Comparisons between the results of the DFRLM and the FRLM indicate that taking driver deviations into account in the model can have a significant effect on the locations chosen and demand covered.

Optimal Spatial Deployment of CO2 Capture and Storage Given a Price on Carbon

Carbon dioxide capture and storage (CCS) links together technologies that separate carbon dioxide (CO2) from fixed point source emissions and transport it by pipeline to geologic reservoirs into which it is injected underground for long-term containment. Previously, models have been developed to minimize the cost of a CCS infrastructure network that captures a given amount of CO2. The CCS process can be costly, however, and large-scale implementation by industry will require government regulations and economic incentives. The incentives can price CO2 emissions through a tax or a cap-and-trade system. This paper extends the earlier mixed-integer linear programming model to endogenously determine the optimal quantity of CO2 to capture and optimize the various components of a CCS infrastructure network, given the price per tonne to emit CO2 into the atmosphere. The spatial decision support system first generates a candidate pipeline network and then minimizes the total cost of capturing, transporting, storing, or emitting CO2. To illustrate how the new model based on CO2 prices works, it is applied to a case study of CO2 sources, reservoirs, and candidate pipeline links and diameters in California.

A Multiobjective Model For Locating Solid Waste Transfer Facilities Using An Empirical Opposition Function

AbstractSolid waste transfer stations (SWTS) are facilities where municipal refuse is transferred from collection trucks to long-haul trucks for more economical shipping to distant landfills. In this paper, a multiobjective model for locating SWTS examines the tradeoffs between minimizing costs and public opposition. The cost objective combines the transshipment and the fixed-charge problems, while expected public opposition is modeled as a decreasing function of distance from the facility. We believe this is the first location model for any type of undesirable facility to use an opposition function derived empirically from opinion survey data (Rahman, et al., 1992). A case study of Phoenix, Arizona uses actual data on sites, zones, tonnages, costs, and local residential attitudes. The model is calibrated to historical budget data for accuracy and also to explore the model’s sensitivity to various parameters. Six sets of multiobjective analyses generate noninferior tradeoff curves under various assumption...