Deep Reinforcement Learning Based Optimization Algorithm for Permutation Flow-Shop Scheduling

Pan, Zixiao; Wang, Jing-jing; Lu, Jiawen

doi:10.1109/tetci.2021.3098354

Cited by 54 publications

(25 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, as seen in [29], there are exact methods [30] that can solve large size instances in only a few minutes. Similarly, Pan et al [31] propose an RL framework that tackles the permutation flow shop scheduling problem. The authors compare the proposal with classical heuristics and an improved versions of these.…”

Section: A Performance Analysismentioning

confidence: 99%

Neural Combinatorial Optimization: a New Player in the Field

Garmendia¹,

Ceberio²,

Mendiburu³

2022

Preprint

View full text Add to dashboard Cite

Neural Combinatorial Optimization attempts to learn good heuristics for solving a set of problems using Neural Network models and Reinforcement Learning. Recently, its good performance has encouraged many practitioners to develop neural architectures for a wide variety of combinatorial problems. However, the incorporation of such algorithms in the conventional optimization framework has raised many questions related to their performance and the experimental comparison with other methods such as exact algorithms, heuristics and metaheuristics. This paper presents a critical analysis on the incorporation of algorithms based on neural networks into the classical combinatorial optimization framework. Subsequently, a comprehensive study is carried out to analyse the fundamental aspects of such algorithms, including performance, transferability, computational cost and generalization to larger-sized instances. To that end, we select the Linear Ordering Problem as a case of study, an NP-hard problem, and develop a Neural Combinatorial Optimization model to optimize it. Finally, we discuss how the analysed aspects apply to a general learning framework, and suggest new directions for future work in the area of Neural Combinatorial Optimization algorithms.

show abstract

Section: A Performance Analysismentioning

confidence: 99%

Neural Combinatorial Optimization: a New Player in the Field

Garmendia¹,

Ceberio²,

Mendiburu³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…However, this method focused on the scale of less than 100 jobs and performs poorly in large-scale problems. Pan et al applied PN [17] in solving the PFSS and explore the actor-critic to train the PFSS models [21], which achieved the state-of-the-art performances for solving the PFSS. However, those RL-based methods take a long time for the network training until the convergence, and the solution accuracy still needs more improvements.…”

Section: Related Workmentioning

confidence: 99%

“…We follow the encoder-decoder [39] learning structure of previous work for the PFSS problems [21]. Instead of using RNN as the encoder [21], we incorporate the graph structure for obtaining better job feature representations. The comparison between the recurrence-based and our graph-based encoders is given in Figure 3.…”

Section: B Policy Networkmentioning

confidence: 99%

“…As for the decoder, we follow the previous work and also utilize the attention mechanism for the decoder [21]. Motivations of using graph encoder.…”

Section: B Policy Networkmentioning

confidence: 99%

“…Thus the learning-based methods attract more attention. However, the state-of-the-art RL method [21] requires heavy networks (an actor network and a critic network) and takes a long time for training until the network converges. To that end, we train the model via expert-driven imitation learning (IL) [22], [23], which leads to more stable and accurate convergences.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Learning to Optimize Permutation Flow Shop Scheduling via Graph-based Imitation Learning

Li¹,

Liang²,

Zhu³

et al. 2022

Preprint

View full text Add to dashboard Cite

The permutation flow shop scheduling (PFSS), aiming at finding the optimal permutation of jobs, is widely used in manufacturing systems. When solving the large-scale PFSS problems, traditional optimization algorithms such as heuristics could hardly meet the demands of both solution accuracy and computational efficiency. Thus learning-based methods have recently garnered more attention. Some work attempts to solve the problems by reinforcement learning methods, which suffer from slow convergence issues during training and are still not accurate enough regarding the solutions. To that end, we train the model via expert-driven imitation learning, which accelerates the convergence more stably and accurately. Moreover, in order to extract better feature representations of input jobs, we incorporate the graph structure as the encoder. The extensive experiments reveal that our proposed model obtains significant promotion and presents excellent generalizability in large-scale problems with up to 1000 jobs. Compared to the state-ofthe-art reinforcement learning method, our model's network parameters are reduced to only 37% of theirs, and the solution gap of our model towards the expert solutions decreases from 6.8% to 1.3% on average. The source code is available at: https://github.com/longkang1/IL_PFSS

show abstract