Rail Intermodal Operation - Coursework Example

Rail Intermodal Operation Intermodal transportation is the backbone of today’s world trade. The purpose of intermodaltransportation is to integrate various forms of transportation with n aim of improving services and increasing efficiency in the distribution process. This mode of transportation is made up of a combination of ship, truck and or rail to transport freight. This paper conducts a research on rail intermodal operation for long-haul. Over the years, rail transport has proved efficient in terms of labor and operation costs. In most cases, rail transport services have diverted freight traffic, reduced traffic congestion, and wear and tear of roads. There have been recent advances in efficiency of intermodal operations, challenges still remain. Some of these challenges result from poor infrastructure, lack of track, and fully operational transcontinental railroad. Other challenges crop up from limitations in management and information. This results to poor train routes, schedules, and rules for programming shipments. The extra delays caused and persistent mishandling of containers at intermodal terminals may make intermodal transportation less efficient (Aggarwal, Oblak and Vermuganti, pg 1079). Scheduling and coordination of trains is improved when the process of scheduling direct and indirect trains is followed concerning the containers to be sent per train for line-haul portion of the intermodal trip. Such a program in the long run, minimizes operation costs. Intermodal rail operations and conventional rail operations The two rail operations differ in several aspects. Intermodal operation networks have few and widely spaced terminal in comparison to the conventional rail operations. Once the networks for intermodal rail terminals get few, economies of scale is realized in container handling as well as train movements within the terminals. Trucks or regional railroads handle the intermodal terminal movements. Another advantage of intermodal rail operations over conventional rail operations is the swiftness in the transfer of containers between trains and few stops made in the journey. This avoids the use of blocks needed to travel in the journey. This in turn reduces train reassembly required in decision making in conventional train scheduling and routing (Bussieck Michael, R., Peter Kreuzer, and Uwe T. Zimmerman). Intermodal rail operations promise shorter delivery lead times when compared to the conventional rail operations. As a result, the need for scheduling trains to increase efficiency and meet customers’ demands. According to this discussion, the first two factors reduce the logistics behind decisions made for intermodal operations over conventional freight. On the other hand, the third factor increases the use of careful train scheduling and decisions made for routes. In most cases, train scheduling and routing relies on demand rates and the aim of attaining steady state train frequencies. Problem Statement This research was facilitated by train-scheduling and problems related to container routing observed at intermodal division of major railroads on the west coast of the US. It is a single hub located in the west-central part with a few intermodal terminals in the east of Mississippi river. A railroad has several intermodal terminals with a flow of traffic on the bounds. The eastern side is dedicated t moving sea cargo for important international shipping lines. The other train capacity is use for utilizing services to smaller customers through few hours of rail road intermodal terminals. The customers use intermodal retailers to connect with truck and rail movements. The retailers reserve space on trains and sell the space to customers later. The reservations create room for predicting demand for the railroad that is exhibited on weekly patterns. This happens due to several factors such as traditional cycles in retail demand, agricultural, as well as manufacturing production. The railroad also provides different levels of service, promising premium delivery. This means that trains may be sent directly from an intermodal terminal to the destination without unnecessary stops at the hub. As a result, it provides the fastest service. Alternatively, trains carrying containers destined for several places may be sent to a hub. At this point, the containers are consolidated in the context of their destination to outbound trains. This process causes delay through transfer or repositioning of rail cars. Further delays may result from poorly coordinated inbound outbound schedules. On the other hand, the capacity of each train is limited. In this case, capacity is expressed in terms of the number of containers. It is assumed that containers are homogeneous in reference to the capacity of the train. This depends on the power of the locomotives and terrain of the route of the rail upon which the train will travel. In actual sense, decisions made regarding the capacity of the train are made prior to demand forecasts. This is based on the assumption that the capacity of the trains in a particular segment is the same, and reflects the condition in the motivation application (Cynthia and Ratliff, pg. 205). The space available for containers in the yard waiting transfer at the hub is limited in the terminals. This implies that as their number increases, they are stacked higher and densely. The time required to retrieve a container increases, placing complications on material handling equipment. This is also a bottleneck. This model fails to constrain the number of containers stored in the yard but does assess the cost of storage for the containers. According to the observations made in the intermodal terminal operations, the schedules for the train and decisions for container routing are not affected in case the speed of delivery is promised. This is a motivation factor on how to schedule trains and route containers so that on-time delivery and efficient services are delivered. Mathematical formulation Mathematical formulation provides the assumption and motivation used in this model to present the integer program used to utilize the rail segment intermodal transportation. The mathematical; structure is discussed in the context of the nature of the problem to assist in understanding the rationale in the new decomposition approach. The intermodal setting combines the various modes of transport. However, this model focuses on rail operations with the assumption that there is enough drayage capacity at the origins as well as the destinations. This is chosen as the focus since the potential for increasing timelines in the intermodal journey through on-time performance of the rail segment is inevitable. In this context, the time period taken is referred to in terms of days. It is assumed that transit and hub delay times are determined and remain constant across time. The transit times and delays at the hub are termed as an integral figure in relation to time periods. In a real life situation, transit times cannot be predicted but since it is expressed in days, slackness in scheduling of unforeseen events results (Cynthia and Ratliff, pg. 206). A rail company incurs operational costs of fixed-charge (setup cost) and variable cost (per unit cost). Fixed charge cost entails wages of operators and cost of assembling the train. The assumption made is that the equipment used is available at the place of need. Another assumption made is the value of the capacity of the train in each transportation segment. These assumptions reflect a reasonable accurate situation for the study. The set up cost for every train in the transportation segment is also assumed, with unlimited number of trains sent per day. The first assumption is slightly logical because the cost of train operator makes a large portion in the set up cost so long as the enough labor is provided. At the same time, the capacity of the train may be limited depending on location and time. This creates some challenges when scheduling the trains. Variable costs consider transportations costs (fuel, oil, and maintenance of the truck), the cost incurred when handling containers on and off the rail cars, and cost incurred in the yard for storage. It is also assumed that transportation costs remain constant regardless of the time, and only depends on origin-destination. This assumption is logical in a short horizon like a week. The other assumption that costs varies on the origin-destination is consistent with the assumption of homogeneous containers used in transportation. Inventory costs are made of costs incurred as a result of yard storage and the chance of possessing unavailable container for use. This is so because the opportunity cost for in-transit goods in incurred by the shipper or the consignee. In this model, yard storage cost is assumed to be equal for all containers, and never varies regardless of ye location and time. It is assumed that clients accept early delivery and so no inventory is incurred at destinations. The model intends to assist in scheduling train operations in a bottleneck direction. The mathematical formula consists inventory holding cost at the origin, as well as the hub, direct and indirect transportation cost of shipped goods, the handling costs (direct and indirect shipment at origin), and the set up cost incurred at the origin for both trains (direct and bound trains). The set up cost for the indirect trains is also factored. The analysis done relies on the assumption of only one hub but the formulation is for general cases that may have several hubs. It is also considered that a container passes through a hub. The direct travel time between the origin and destination is assumed to be less than the total transit and delay time for indirectly shipped container. Network description and formulation The problem can be represented in a network describing two proportional representations: single-commodity network and multi-commodity network. The single-commodity network shows only one type of flow on each arc. The multi-commodity network shows several commodities flowing on each arc. The commodities are defines by their origin, destination, arrival date at the origin as well as the destination due date. The single commodity has a set of nodes and arcs that represent physical location of the commodity at a particular point as each arc bears the flow of the single commodity in the context of time and space. The travel time in a single-commodity network is assumed to be direct. Handling time at the hub is also assumed to be direct. This implies that the time indices for the flow and arcs reflect the real transit and times of delay. An n important observation made is that the super source and super sink has three types of nodes; the set of nodes per triple, a set of nodes per triple, and another set of nodes per triple. The triple consists of the following nodes; destination, time period, and commodity. The arcs from the super source to nodes have upper a lower capacity limits that are equal to the number of containers arriving at the destination and time. In the first set, the arcs and nodes to the third replicates direct shipment of the container, as the arcs from the first nodes in set three pass through nodes in the second enroute. The containers flow in a super sink manner from each node in set three whose arrival and due dates coincide. Inventory arcs connect an origin and a hub for a particular commodity to the corresponding node. The container flows in this network shows locations of storage of containers from one period to another. Handling and transportation costs are monitored between the origins, hubs, and destinations, as inventory holding costs are monitored along the arcs but between sequential time periods. At the super source and super sink, no cost emanates. This network factors in train costs and capacity constraints (Aggarwal, Oblak and Vermuganti, pg. 1078). The multi-commodity network is a simple version of the single commodity network. This is so because the nodes are not highlighted by the type of commodity. This implies that the nodes are differentiated by location and period. The flows correspond to the feasible movement to the container between the two stations over a given period of time. At times, it depends on the inventory storage data in a particular station from one period of time to another. In this network, different commodities with similar time and locations flow simultaneously on the same arc. Setup costs are done depending on the number of trains assessed and the flow of the containers in the single arc. Despite the type of network used, train costs are monitored, as capacity constraints get imposed on the related flow for multiple commodity networks. This implies that the multistep multicommodity flow experiences problems. Classical Decomposition Techniques Due to the problems raised from the multi-setup mulicommodity network, two classical decomposition procedures are applied; lagrangian relaxation and Bender’s Decomposition. The Lagrangian Relaxation is a procedure applied to assist solve difficult integer and mixed integer programs. The procedure does not solve the problem directly but looks for probable multipliers through sub gradient optimization. The Lagrangian applies another approach of relaxing part or all constraints in the network but not the capacity constraints of the train. As a result, container routing decisions are made, with a focus on the constraints of the capacity of the train (Bussieck, Kreuzer, and Zimmerman, pg.54-60). Bender’s Decomposition While the Lagrangian technique treats troublesome constraints specifically, the Bender’s Decomposition controls the influence of troublesome variables. The problem is divided into a master problem (integer program) and the sub problem (linear program). The reason behind this division is the simplicity by which liner program is solved once the integer variables calculated from the master problem are established. The purpose of the master problem is to minimize the subject to low bound constraints in a single scalar form. The low bound constraints are defined by expressions differing between the terms with integer variables with original objectives. The product derived on the right hand side of the constraints and the dual multipliers are the sub problem. This sub problem has the same form of as the original problem although the troublesome variables remain specific. This technique connects the master problem with the sub problem and is based on current values of integer variables. During each iteration new dual multipliers are generated. The dual multipliers give information on the values of the variables in the integer program. The values can be changed to improve the solution for the original problem. However, this technique fails to offer viable solutions to the problems. Decomposition technique This approach is motivated by the optimal pattern through which the containers arrive. This is looked at in the context of origin, destination, arrival and the due date of the container. When this is determined, it is possible to infer the containers that require direct trains. However, for all pattern of arrival of the containers, three problems exist; scheduling of direct trains per destination pair, scheduling the trains and containers in the hub according to origin, and scheduling the trains and containers from the hub according to their destination. These problems can be decomposed into two sections; scheduling trains and containers to provide maximum shipment directly for their bound destinations or hub, and optimal scheduling of trains and containers (Bussieck, Kreuzer, and Zimmerman, 60-63) Conclusion This paper has discussed the near-optimal solutions to train scheduling and decision making process for rail intermodal setting. The paper highlights twofold approach to intermodal rail transportation. First, is the formulation developed in the setting and the advantages in the network structure. Secondly, the solution procedure through the available techniques. The solutions bear near-optimal solutions to problems. Works Cited Aggarwal, A. K., Oblak and R.R. Vermuganti. "A Heuristic Solution Procedure for Multicommodity Integer Flows." Computersand Operations Research, Vol. 22, No. 10 (1995): 1075-1087. Print. Barnhart Cynthia and H. Donald Ratliff. "Modelling Intermodal Rotuing." Journal of Business logistics, Vol. 14, No. 1. (1993): 205-223. print. Bussieck Michael, R., Peter Kreuzer, and Uwe T. Zimmerman. "Optimal Lines for Railway Systems." European Journal of Operational research, Vol. 96, No. 1 (1997): 54-63. Print. Read More