Abstract Rail freight planning problems pose specific challenges that have attracted the attention of academics and industry professionals for many decades. They involve multiple types of assets (trains, stations, terminals, etc.) and are subjected to structural, operational and safety constraints. Even though various approaches have been proposed, few can address the complexity and size of real‐world scenarios, and decentralized techniques, like multi‐agent systems (MAS), have become more prevalent. The current state of the art in disciplines such as agent technology, reinforcement learning and discrete‐event simulation allows the implementation of complex architectures, with multiple actors interacting and learning simultaneously. Therefore, this study takes advantage of these current advances and proposes an innovative approach to real‐time traffic management problems in freight railway networks through multi‐agent deep reinforcement learning (MADRL). This study was motivated by the decision‐making scheduling problems arising in the Hunter Valley Coal Chain (HVCC), located in New South Wales, Australia. The MADRL algorithm uses as the training environment the simulation model currently utilized for capacity planning of the HVCC, allowing experiments with actual data. Thus, we enhanced the simulation model to accommodate a MAS with intelligent agents representing system elements, such as trains, dump stations, and load points. Furthermore, these agents act in a decentralized fashion based on local observations, constituting a partially‐observed Markov decision process (dec‐POMDP). Three variations of the MADRL approach are presented: a baseline model, an extended model, and one that directly addresses deadlocks. Finally, we present a transfer learning method that improves deadlock resolution and leverages performance. In the experiments, we explore specific, complex scenarios arising in the HVCC, where trains frequently face deadlock conditions. The baseline model outperforms a first‐come‐first‐serve (FCFS) based heuristic used by HVCC's simulation model and a genetic algorithm in instances with up to 60 trains – but fails in more complex scenarios. On the other hand, the most advanced model, which addresses deadlocks via transfer learning, always finds feasible solutions and produces policies that outperform the FCFS‐based heuristic in 94% of the instances.

Abstract. This note corrects an error in our paper “N. Boland, R. Clement, and H. Waterer. A bucket indexed formulation for nonpreemptive single machine scheduling problems. INFORMS Journal on Computing 28(1):14–30, 2016.”

The Hunter Valley coal chain, located in New South Wales, Australia, is one of the most complex supply chains in the world. Coal orders are moved from the mines in the region to the terminals using a specific, complex rail infrastructure. These operations are scheduled by an experienced planning team at the Hunter Valley Coal Chain Coordinator. In this study, we propose an improved Genetic Algorithm to address the train scheduling problem. Our model considers several real-life operational constraints present in the coal supply chain and includes the selection of trains from an available fleet. Using a rail network with most of the real Hunter Valley railway infrastructure, we evaluate the strategy on test instances generated from actual train operations between 2017 and 2018. The objective of our strategy is to minimize total travel times. The algorithm was evaluated on instances with sizes between 60 and 180 jobs, and results show that the method can reach high-quality solutions - i.e. similar or better than those being currently used-in less than 2 minutes for the smaller instances, and 20 minutes for the larger ones.

A new mixed-integer linear programming (MILP) formulation for nonpreemptive single machine scheduling problems is presented. The model is a generalisation of the classical time indexed (TI) model to one in which at most two jobs can be processing in each time period. Like the TI model, the new model, called the bucket indexed (BI) model, partitions the planning horizon into periods of equal length, or buckets. Unlike the TI model, the length of a period is a parameter of the BI model and can be chosen to be as long as the processing time of the shortest job. The two models are equivalent if a period is of unit length, but when longer periods are used in the BI model, it can have significantly fewer variables and nonzeros than the corresponding TI model. A computational study using weighted tardiness instances, and weighted completion time instances with release dates, reveals that the BI model significantly outperforms the TI model on instances where the minimum processing time of the jobs is large. Furthermore, the performance of the BI model is less vulnerable to increases in average processing time when the ratio of the largest processing time to the smallest is held constant.

This study is part of a project which investigates computational principles which underlie perception and representation of architectural streetscape character. Some of the principles can be associated with fundamental concepts in brain theory and Gestalt psychology. For the experimental analysis streetscapes were represented by sequences of digital images of house facades which were prepared by a team of researchers from architecture. Two methods for non-linear dimensionality reduction, isomap and maximum variance unfolding, were applied to a set of Hough arrays (for lines) of the given images. An analysis of the extracted "streetmanifolds" revealed groupings of house facades with similar visual character and proportions. Comparative tests were conducted on a simple cylinder shaped example manifold to evaluate the geometric stability of the two dimensionality reduction methods. All experiments addressed variations of the distance metric and the neighbourhood parameter