Paul Corry

An assignment model for dynamic load planning of intermodal trains

Intermodal terminals are important facilities in the container transport network, providing an exchange of containers between road and rail transport. Numerous factors can affect throughput in such highly integrated systems. These include numbers and types of equipment, physical layout, storage capacity and operating strategies. This study aims to improve operating strategies by developing an analytical tool to assist in load planning of container trains. The problem investigated can be described as a dynamic assignment problem with many uncertain parameters. Numerical investigations focus on tuning the proposed model to deal with the uncertainties.

Ant Colony Optimisation for Machine Layout Problems

Flexible machine layout problems describe the dynamic arrangement of machines to optimise the trade-off between material handling and rearrangement costs under changing and uncertain production environments. A previous study used integer-programming techniques to solve heuristically reduced versions of the problem. As an alternative, this paper introduces an ant colony optimisation (ACO) algorithm to generate good solutions. Experimental results are presented, with ACO obtaining better solutions than the reduction heuristic.

Optimised loading patterns for intermodal trains

This paper considers one important aspect of operations planning referred to hereafter as train planning. Train planning is the process of spatially assigning containers to specific wagons (also known as railcars) on an intermodal train. The spatial arrangement of containers on a train can have a significant influence over the amount of time and energy consumed in the handling of containers. Efficient train planning can also maximise utilisation of wagon carrying capacity. This study proposes a mixed-integer programming model to determine the arrangement of containers on a train to minimise a weighted sum of number of wagons required and equipment working time. Due to the large number of variables, the proposed model cannot be solved in a timely manner for practical problems. This is addressed by applying heuristic algorithms local search and simulated annealing. Discrete-event simulation of an intermodal terminal is used to evaluate the proposed methods and to illuminate various properties of the model.

Job scheduling with technical constraints

Many scheduling problems that arise in industry have technical constraints unique to the specific industry. Scheduling methodologies must be highly customized to deal with the unique technical constraints. This study proposes a scheduling model that can incorporate technical constraints into standard scheduling constraints already present in classical models. Using this approach, technical constraints from one industry can be interchanged with those from another with little modification to the existing methodologies. The conditions under which this approach can be applied are investigated and frameworks for applying dispatching rules are proposed. Numerical experiments evaluate the performance of these dispatching rules and compare them with two meta-heuristics.

Meta-Heuristics for a Complex Push-Pull Production System

Hybrid push/pull production systems have received recent attention in the literature. This paper investigates a hybrid push/pull system originating from a foundry. The aim is to determine safety stock and replenishment levels for a large inventory situated at the junction point between component production and assembly operations. Components are produced according to a make-to-stock policy and are received into inventory when completed. Assembled goods are made-to-order, pulling components from the inventory when required. Classical techniques cannot be used in such a complex environment because they are based on invalid assumptions. This study proposes heuristically controlled simulations for attaining good solutions to the problem. Experimental results demonstrate and compare the proposed methods.

Optimising a Horizontally Integrated Push/Pull Hybrid Production System in a Foundry

This study investigates a horizontally integrated push/pull hybrid production system (HIHPS) of a large foundry in the Asia-Pacific region. Components are produced according to a pushing philosophy and are stored in a buffer when complete. Assembled goods draw components from the buffer and are made to order. The objective is to minimise the sum of ordering, inventory and tardiness costs. Currently, the inventory decisions are made based on the experience of personnel within the foundry. Operations research provides an opportunity to improve efficiency by using computational techniques to determine good inventory decisions. Traditional inventory and lot-sizing problems determine optimal policies using mathematical analysis made possible by simplifying assumptions. These assumptions are invalid for the HIHPS problem studied in this paper so that mathematical analysis is extremely difficult if not impossible. As an alternative, simulation is combined with simulated annealing to determine a good inventory policy. Experimental results are used to evaluate the performance of this approach and to identify strategies for developing even better techniques.

Integrating Ship Scheduling and Berth Allocation for Container Seaports with Channel Access

In this talk we discuss the discrete berth allocation problem under channel access restrictions given by a river or sea gate through which calling vessels have to transit. We further outline an optimization concept for scheduling vessel movements in the channel related with the well known flowshop machine scheduling problem.