A review of Pareto pruning methods for multi-objective optimization

Petchrompo, Sanyapong; Coit, David W.; Brintrup, Alexandra; Wannakrairot, Anupong; Parlikad, Ajith Kumar

doi:10.1016/j.cie.2022.108022

Cited by 93 publications

(26 citation statements)

References 150 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In the multi-objective optimization problems, the objective functions are contradictory to each other and cannot be optimal at the same time, so the final result is a Pareto-optimal set of solutions. In practical engineering, it is often necessary to select a solution from the Pareto-optimal set of solutions based on three indicators (the cost, the wind and PV power consumption rate, and the loss of load probability) as the model optimization result [53].…”

Section: Improved Multi-objective Genetic Algorithm 31 Basic Theory O...mentioning

confidence: 99%

An Optimization Capacity Design Method of Wind/Photovoltaic/Hydrogen Storage Power System Based on PSO-NSGA-II

Liu¹,

Liu²

2023

Energy Engineering

View full text Add to dashboard Cite

The optimal allocation of integrated energy system capacity based on the heuristic algorithms can reduce economic costs and achieve maximum consumption of renewable energy, which has attracted many attentions. However, the optimization results of heuristic algorithms are usually influenced by the choice of hyperparameters. To solve the above problem, the particle swarm algorithm is introduced to find the optimal hyperparameters of the heuristic algorithms. Firstly, an integrated energy system consisting of the photovoltaic, wind turbine, electrolysis cell, hydrogen storage tank, and energy storage is established. Meanwhile, the minimum economic cost, the maximum wind and PV power consumption rate, and the minimum load shortage rate are considered to be the objective functions. Then, a hybrid method combined the particle swarm combined with non-dominated sorting genetic algorithms-II is proposed to solve the optimal allocation problem. According to the optimal result, the economic cost is 6.3 million RMB, and the load shortage rate is 9.83%. Finally, four comparative experiments are conducted to verify the superiority-seeking ability of the proposed method. The comparative results indicate that the proposed method possesses a stronger merit-seeking ability, resulting in a solution satisfaction rate of 87.37%, which is higher than that of the unimproved non-dominated sorting genetic algorithms-II.

show abstract

Section: Improved Multi-objective Genetic Algorithm 31 Basic Theory O...mentioning

confidence: 99%

An Optimization Capacity Design Method of Wind/Photovoltaic/Hydrogen Storage Power System Based on PSO-NSGA-II

Liu¹,

Liu²

2023

Energy Engineering

View full text Add to dashboard Cite

show abstract

“…Many approaches have been proposed to select an optimal or sub-optimal ensemble of traditional ML classifiers [1][2][3][4]. Recently, due to the inherent model-size issues in deep learning, network compression techniques have emerged as a new and challenging area to alleviate the problem of rapidly increasing memory and computational requirements.…”

Section: Introductionmentioning

confidence: 99%

Compression of Deep-Learning Models Through Global Weight Pruning Using Alternating Direction Method of Multipliers

Lee

Hwangbo

Yang

et al. 2023

Int J Comput Intell Syst

View full text Add to dashboard Cite

Deep learning has shown excellent performance in numerous machine-learning tasks, but one practical obstacle in deep learning is that the amount of computation and required memory is huge. Model compression, especially in deep learning, is very useful because it saves memory and reduces storage size while maintaining model performance. Model compression in a layered network structure aims to reduce the number of edges by pruning weights that are deemed unnecessary during the calculation. However, existing weight pruning methods perform a layer-by-layer reduction, which requires a predefined removal-ratio constraint for each layer. Layer-by-layer removal ratios must be structurally specified depending on the task, causing a sharp increase in the training time due to a large number of tuning parameters. Thus, such a layer-by-layer strategy is hardly feasible for deep layered models. Our proposed method aims to perform weight pruning in a deep layered network, while producing similar performance, by setting a global removal ratio for the entire model without prior knowledge of the structural characteristics. Our experiments with the proposed method show reliable and high-quality performance, obviating layer-by-layer removal ratios. Furthermore, experiments with increasing layers yield a pattern in the pruned weights that could provide an insight into the layers’ structural importance. The experiment with the LeNet-5 model using MNIST data results in a higher compression ratio of 98.8% for the proposed method, outperforming existing pruning algorithms. In the Resnet-56 experiment, the performance change according to removal ratios of 10–90% is investigated, and a higher removal ratio is achieved compared to other tested models. We also demonstrate the effectiveness of the proposed method with YOLOv4, a real-life object-detection model requiring substantial computation.

show abstract

“…Our approach also offers data owners and other stakeholders a framework to guide them in the selection of the anonymization strategy to apply, by facilitating understanding of the trade-off between privacy and data utility. To this end, we employ a multi-objective optimization method based on Pareto optimality [34] to assist data owners in selecting optimal anonymization strategies according to their privacy and data utility requirements.…”

Section: Introductionmentioning

confidence: 99%

A decision-support framework for data anonymization with application to machine learning processes

Caruccio

Desiato²,

Polese³

et al. 2022

Information Sciences

View full text Add to dashboard Cite

A review of Pareto pruning methods for multi-objective optimization

Cited by 93 publications

References 150 publications

An Optimization Capacity Design Method of Wind/Photovoltaic/Hydrogen Storage Power System Based on PSO-NSGA-II

An Optimization Capacity Design Method of Wind/Photovoltaic/Hydrogen Storage Power System Based on PSO-NSGA-II

Compression of Deep-Learning Models Through Global Weight Pruning Using Alternating Direction Method of Multipliers

A decision-support framework for data anonymization with application to machine learning processes

Contact Info

Product

Resources

About