The recent end-to-end neural solvers have shown promise for small-scale routing problems but suffered from limited real-time scaling-up performance. This paper proposes GLOP (Global and Local Optimization Policies), a unified hierarchical framework that efficiently scales toward large-scale routing problems. GLOP hierarchically partitions large routing problems into Travelling Salesman Problems (TSPs) and TSPs into Shortest Hamiltonian Path Problems. For the first time, we hybridize non-autoregressive neural heuristics for coarse-grained problem partitions and autoregressive neural heuristics for fine-grained route constructions, leveraging the scalability of the former and the meticulousness of the latter. Experimental results show that GLOP achieves competitive and state-of-the-art real-time performance on large-scale routing problems, including TSP, ATSP, CVRP, and PCTSP. Our code is available at: https://github.com/henry-yeh/GLOP.

Combustion optimization of hot blast stoves is a promising approach for cost savings and energy conservation of ironmaking. Existing artificial intelligence methods for this optimization rely on air and gas flow meters, which can malfunction under harsh working conditions. To meet this challenge, we propose an intelligent combustion control system based on reinforcement learning (RL). Considering the difficulty of learning state feature representation, five RL models using different deep embedding networks were implemented and evaluated. The Attention-MLP-based RL model is distinguished through experimental testing, achieving an accuracy of 85.91% and an average inference time of 4.85 ms. Finally, the intelligent combustion control system with the Attention-MLP-based RL model runs in the hot blast stove of the blast furnace (1750 m³ in volume) at Tranvic Steel Co., Ltd. in China (Chengdu, China). The results show that our system can achieve good control performance by autonomously learning the implicit relationship between the state of the hot blast stove and the valve control action in the absence of flow meters.

Generally, the result of clustering cannot reflect the similarities of time series properly because of the disturbance of noises and details in time series. In this paper, a new approach to this problem based on wavelet decomposition and denoising is proposed. The approach has been tested and analyzed by Synthetic Control Chart Time Series from University of California,Irvine (UCI) knowledge discovery in databases (KDD).

Blast furnace (BF) ironmaking requires flexible and precise burden distribution. As a critical distributing unit, the angle diagnosis of BF chute is essential for operation under severe conditions. However, the existing methods must be performed during BF overhauls, which cannot satisfy timely and convenient diagnosis requirements. Inspired by cutting-edge computer vision technologies, a novel angle diagnosis method for BF chute based on deep learning of temporal images is proposed, which uses grayscale consistency transformation and data augmentation for data pre-processes, and a Res-LSTM based deep neural network model for angle diagnosis. Experiments were conducted using chute motion videos collected from the BF top imaging system, and experimental results showed that our method can extract spatial-temporal features and accurately identify chute angles. The BF chute angle diagnosis system developed based on this method has been successfully running in No.7 BF (1750m ³ in volume) of Tranvic Steel Co., Ltd in China for half a year. The system can accurately diagnose the chute angle in real time, for a diagnosis rate of 100%.

Ant Colony Optimization (ACO) is a meta-heuristic algorithm that has been successfully applied to various Combinatorial Optimization Problems (COPs). Traditionally, customizing ACO for a specific problem requires the expert design of knowledge-driven heuristics. In this paper, we propose DeepACO, a generic framework that leverages deep reinforcement learning to automate heuristic designs. DeepACO serves to strengthen the heuristic measures of existing ACO algorithms and dispense with laborious manual design in future ACO applications. As a neural-enhanced meta-heuristic, DeepACO consistently outperforms its ACO counterparts on eight COPs using a single neural architecture and a single set of hyperparameters. As a Neural Combinatorial Optimization method, DeepACO performs better than or on par with problem-specific methods on canonical routing problems. Our code is publicly available at https://github.com/henry-yeh/DeepACO.