Concordia University

This paper studies the online vector bin packing (OVBP) problem and the related problem of online hypergraph coloring (OHC). Firstly, we use a double counting argument to prove an upper bound of the competitive ratio of $FirstFit$ for OVBP. Our proof is conceptually simple, and strengthens the result in Azar et. al. by removing the dependency on the bin size parameter. Secondly, we introduce a notion of an online incidence matrix that is defined for every instance of OHC. Using this notion, we provide a reduction from OHC to OVBP, which allows us to carry known lower bounds of the competitive ratio of algorithms for OHC to OVBP. Our approach significantly simplifies the previous argument from Azar et. al. that relied on using intricate graph structures. In addition, we slightly improve their lower bounds. Lastly, we establish a tight bound of the competitive ratio of algorithms for OHC, where input is restricted to be a hypertree, thus resolving a conjecture in Nagy-Gyorgy et. al. The crux of this proof lies in solving a certain combinatorial partition problem about multi-family of subsets, which might be of independent interest.

Reinforcement learning has recently emerged as a promising approach for supervisory control and operational optimization in complex industrial systems. In this work, we propose a hierarchical learning-based framework for the energy-efficient real-time operation of a photovoltaic reverse osmosis (PVRO) desalination plant under time-varying freshwater demand. The proposed framework combines a high-level reinforcement learning (RL)-based planning agent, implemented using Proximal Policy Optimization (PPO), with a lowlevel proportional-integral-derivative (PID) controller to jointly address operational efficiency and control stability. Using a dynamic model of the RO process, the PPO agent determines permeate flow-rate set-points by accounting for water demand, storage-tank dynamics, energy consumption, and operational constraints. The selected set-points are tracked by the PID controller, which regulates permeate flow by adjusting the feed pressure of the high-pressure pump, ensuring accurate and stable flow control. The reward formulation penalizes excessive energy use and unsafe storage conditions, including tank overflow and underflow, while promoting demand satisfaction. Simulation results demonstrate effective real-time PVRO operation while maintaining safe storage levels and reliable freshwater production. Under grid-only operation, the PPO-based planner reduces daily grid-energy consumption by approximately 4%-5% compared with rule-based scheduling baselines and achieves near-oracle performance, differing from the day-ahead oracle benchmark by only 17 kWh/day. Under PV-assisted operation, the proposed framework reduces grid-energy consumption from 6754 to 4677 kWh/day, corresponding to a 30.7% reduction relative to grid-only operation. The separation of planning and control layers provides a practical and computationally efficient solution that can be integrated with existing industrial RO control systems.

Lithium-rich layered oxides (LRLOs) are regarded as promising cathode materials for next-generation Lithium-ion batteries (LIBs). However, they face challenges such as structural instability, voltage decay and capacity that hinder their practical application and commercialisation. This study presents the first comprehensive bibliometric analysis of global LRLO cathode research covering the period 2005–2025 using 1048 publications retrieved from the Scopus and Web of Science databases. The results show that China leads globally in publication output, institutional contributions, and international collaborations. The lifecycle analysis indicates that LRLO research has evolved into a mature yet rapidly expanding scientific domain, while Lotka's Law confirms the establishment of a stable scientific community. The major research themes include "lithium-ion batteries", "high-capacity", "cathodes", "electrochemical performance", "surface modification", and "oxygen redox chemistry", indicating a shift from materials synthesis to mechanism-based optimisation. These findings provide strategic insights into the current development of LRLO cathodes and highlight AI-assisted materials design as a promising opportunity to accelerate their optimisation and commercialisation.

Rail-transit timetables (e.g., light rail, suburban rail, and metro systems) have traditionally been established through hierarchical, top-down procedures where service frequencies are specified first, based on aggregated demand indicators or institutional targets, and detailed timetables are subsequently constructed to meet these specifications under operational constraints. Recent advances in data availability and modeling capabilities have enabled a shift toward what we term demand-driven timetabling. Unlike hierarchical approaches, demand-driven timetabling strategies directly optimize the timetable while accounting for time-varying passenger flows, thus jointly optimizing service quality and operational efficiency. We develop a detailed taxonomy of the main application scenarios, modeling archetypes and objective functions adopted in the demand-driven timetabling literature, and provide an analysis of the strategies used to represent train capacities and passenger boarding behavior. We also discuss the most prominent modeling features and assumptions, including rolling-stock circulation, dwell and travel time modeling, as well as energy-aware scheduling. Finally, we provide a comprehensive overview of the solution methodologies employed, ranging from exact and decomposition algorithms to metaheuristic and control-oriented approaches. We conclude by identifying key trends, limitations, and research gaps in the field, outlining future directions for adaptive timetable design to support the next generation of passenger-centric and energy-efficient rail-transit systems.

Abstract Water loss in distribution networks represents a critical challenge in urban water management, with significant environmental, economic, and public health implications. Although previous studies developed both leak detection and localization approaches, detection has been prioritized. This paper presents a hybrid deep-learning approach for leakage localization with a proposed loss function that minimizes the distance between predicted and actual leak locations in water distribution networks. A hybrid CNN-LSTM model was developed, combining a Convolutional Neural Network (CNN) for spatial feature extraction across the sensor array and a Long Short-Term Memory (LSTM) network for sequence modeling. In addition, a distance-based loss function was designed to optimize both classification accuracy and the physical distance between predicted and actual leak localizations. The algorithm was validated using the L-Town water distribution network benchmark. Results demonstrate that incorporating network distance into training improves localization accuracy compared to a standard cross-entropy loss. Average localization errors of 181 m for background leakages and 58 m for bursts were achieved with the distance-based loss function approach. These results represent an improvement of about 32% for both leak types relative to the standard loss function. The contribution of this paper is two-fold: first, the accuracy of leakage localization is improved through a hybrid CNN-LSTM model trained with a distance-based loss function; second, the search area for leak localization is reduced, as incorporating network distance into training minimizes the physical distance between predicted and actual leak locations. This reduction in search area translates directly to decreased inspection time and costs for water utilities.