Ling Wang’s research while affiliated with Tsinghua University and other places

Distributed manufacturing and fine-manufacturing are two typical scenarios of modern manufacturing industries in the context of globalization and customization. The distributed differentiation flowshop integrated scheduling problem (DDFISP) is a novel model that deals with the integrated scheduling problem of these two manufacturing scenarios. In the DDFISP, jobs have multiple customized types and are manufactured in a number of distributed factories. Each factory includes three fine-processing stages: parallel machine fabrication, single machine assembly, and dedicated machine differentiation. In the paper, a new distributed memetic evolutionary architecture is first built, which consists of four modules with distinct functions, including global exploration, local exploitation, knowledge transfer, and search restart. The exploration and exploitation are coevolved in the distributed way and communicated by knowledge transfer. This architecture can be used as a universal model to construct evolutionary algorithms. Following this architecture and devising each module innovatively, a novel knowledge transfer-driven distributed memetic algorithm (KTDMA) is constructed to solve the DDFISP. Specifically, global exploration is performed on multiple populations by dynamically selecting global exploration optimizers from predefined external repository. Local exploitation is executed on an independent elite archive by a destruction-construction local search and a key block local search. Knowledge transfer is conducted to communicate the superior information between exploration and exploitation based on a point-ring topology. Search restart is adaptively carried out to alleviate the search homogeneity. Computational results show the effectiveness of the proposed evolutionary architecture and special designs, and demonstrate that the KTDMA performs more competitive than the compared state-of-the-art algorithms.

Compared to existing distributed flowshop scheduling problems (DFSPs), this paper addresses a more realistic DFSP, which integrates intermachine blocking constraints and two customized processing stages of assembly and differentiation. The manufacturing process includes job fabrication in distributed blocking flowshops, job-to-product assembly on an assembling machine, and product fine-processing on differentiation machines. A novel evolutionary framework consisting of continuous space exploration, discrete space exploitation, and dual space knowledge migration is devised. This framework has advanced features of distribution, memetic evolution, and dual-space coevolution, and can serve as a unified model to construct algorithms for different optimization problems. Based on this evolutionary framework, a matrix and learning co-aided distributed dual-space memetic algorithm (DDMA) is proposed to address the studied problem. In DDMA, exploratory population is represented as a real matrix, where individuals have different identities that will dynamically adjust with evolution. In accordance with identity differences, exploratory population is heterogeneously evolved in the continuous search space by a matrix-aided evolutionary optimizer. The exploitative population consists of elite individuals, which are represented as discrete permutations. It is evolved in parallel with exploratory population and in the discrete search space by a learning-aided evolutionary optimizer, including a reinforcement learning-based multi-neighborhood local search and a statistical learning-based enhanced local search. To communicate the superior evolutionary information obtained by exploration and exploitation, an adaptive knowledge migration across continuous and discrete search spaces is proposed based on the impact of migration on the population diversity. The computational results demonstrate the superiority of DDMA over state-of-the-art algorithms.

In expensive multi-objective optimization, where the evaluation budget is strictly limited, selecting promising candidate solutions for expensive fitness evaluations is critical for accelerating convergence and improving algorithmic performance. However, designing an optimization strategy that effectively balances convergence, diversity, and distribution remains a challenge. To tackle this issue, we propose a composite indicator-based evolutionary algorithm (CI-EMO) for expensive multi-objective optimization. In each generation of the optimization process, CI-EMO first employs NSGA-III to explore the solution space based on fitness values predicted by surrogate models, generating a candidate population. Subsequently, we design a novel composite performance indicator to guide the selection of candidates for real fitness evaluation. This indicator simultaneously considers convergence, diversity, and distribution to improve the efficiency of identifying promising candidate solutions, which significantly improves algorithm performance. The composite indicator-based candidate selection strategy is easy to achieve and computes efficiency. Component analysis experiments confirm the effectiveness of each element in the composite performance indicator. Comparative experiments on benchmark problems demonstrate that the proposed algorithm outperforms five state-of-the-art expensive multi-objective optimization algorithms.

Existing methods for natural gas pipeline leak detection primarily rely on feature extraction and fault classification based on Euclidean data, which often overlook the potential spatiotemporal correlations present in leak signals. This limitation restricts the development of high-precision fault identification models. To address this issue, a Spatio-Temporal Data-Driven Framework (STDDF) for acoustic signal-based natural gas pipeline leak detection is proposed. The framework begins with the application of a percentile-based cosine similarity measure to the acoustic sensor data, transforming it into a graph structure that captures both topological and temporal correlations. Subsequently, a full-graph mean aggregation method is introduced, utilizing a graph sample and aggregation network to enhance the representation of leak signal features. This approach overcomes the limitations of single-node features in capturing the spatial complexity of pipeline leaks. Furthermore, a gated recurrent unit is employed to process the time-series data, effectively capturing the temporal dynamics of leak signals. Validation experiments conducted on a pipeline leak detection platform demonstrate that the STDDF model significantly outperforms traditional deep learning methods, achieving a fault identification accuracy of up to 99 %.

—This study addresses the lot-streaming hybrid job shop scheduling problem with variable sublots (LHJSV), inspired by a real-world aircraft tooling shop. A computational model is developed to represent the complex scheduling processes of the tooling shop. To solve this problem, we propose an automatic memetic algorithm enhanced by a heuristic designed with the assistance of a large language model (LLM). The approach is designed as follows: first, a memetic computing framework with automated algorithmic design is proposed for LHJSV. Second, a cooperative evolutionary heuristic framework based on problem decomposition is introduced, enabling the LLM to comprehend the LHJSV characteristics and generate feasible algorithms. Third, problem-specific prompts for LHJSV are carefully designed to guide the LLM. To evaluate the effectiveness of the proposed method, 20 benchmark instances derived from the Taillard dataset and a real-world case involving 575 operations are utilized. The proposed algorithm is compared against three swarm-based algorithms, an end-to-end method, and an LLM-based algorithm. Experimental results demonstrate that our method outperforms the compared algorithms on 85% of benchmark instances and exhibits significant superiority in real-world scenarios.

Deterministic processing time are no longer applicable under realistic circumstances because of the uncertainties involved in manufacturing and production processes. The present study aims to address a multiobjective distributed assembly flexible job shop scheduling problem with type-2 fuzzy time (DAT2FFJSP), focusing on the optimization objectives of minimizing the makespan and total energy consumption. To address this problem, a mixed-integer linear programming model is presented. Then, a population-based iterative greedy algorithm (PBIGA) with a Q -learning mechanism is proposed, which possesses the following characteristics: 1) a hybrid initialization method is used to generate the population; 2) six local search operators, crossover operators, and mutation operators are applied to explore and exploit the solution space; and 3) the Q -learning mechanism intelligently utilizes historical information on the success of local search operator updates to determine the most suitable perturbation operator; and 4) an energy-saving strategy is applied to improve the candidate solutions. Finally, the effectiveness of the proposed components is validated through extensive experiments that are conducted on 30 instances. The PBIGA outperforms the state-of-the-art algorithms on the DAT2FFJSP.

Differential evolution (DE) algorithm is recognized as one of the most effective evolutionary algorithms, demonstrating remarkable efficacy in black-box optimization due to its derivative-free nature. Numerous enhancements to the fundamental DE have been proposed, incorporating innovative mutation strategies and sophisticated parameter tuning techniques to improve performance. However, no single variant has proven universally superior across all problems. To address this challenge, we introduce a novel framework that employs reinforcement learning (RL) to automatically design DE for black-box optimization through meta-learning. RL acts as an advanced meta-optimizer, generating a customized DE configuration that includes an optimal initialization strategy, update rule, and hyperparameters tailored to a specific black-box optimization problem. This process is informed by a detailed analysis of the problem characteristics. In this proof-of-concept study, we utilize a double deep Q-network for implementation, considering a subset of 40 possible strategy combinations and parameter optimizations simultaneously. The framework's performance is evaluated against black-box optimization benchmarks and compared with state-of-the-art algorithms. The experimental results highlight the promising potential of our proposed framework.