Reducing idleness in financial cloud services via multi-objective evolutionary reinforcement learning based load balancer

Science China Information Sciences, Journal Year: 2024, Volume and Issue: 67(2)

Published: Jan. 25, 2024

Language: Английский

---

Large Language Models as Evolutionary Optimizers

2022 IEEE Congress on Evolutionary Computation (CEC), Journal Year: 2024, Volume and Issue: unknown, P. 1 - 8

Published: June 30, 2024

Language: Английский

---

Learn to optimize—a brief overview

National Science Review, Journal Year: 2024, Volume and Issue: 11(8)

Published: April 2, 2024

Most optimization problems of practical significance are typically solved by highly configurable parameterized algorithms. To achieve the best performance on a problem instance, trial-and-error configuration process is required, which very costly and even prohibitive for that already computationally intensive, e.g. associated with machine learning tasks. In past decades, many studies have been conducted to accelerate tedious from set training instances. This article refers these as

Language: Английский

---

Automated design of state transition rules in ant colony optimization by genetic programming: a comprehensive investigation

Memetic Computing, Journal Year: 2025, Volume and Issue: 17(1)

Published: Jan. 28, 2025

Language: Английский

---

A Population Cooperation based Particle Swarm Optimization algorithm for large-scale multi-objective optimization

Swarm and Evolutionary Computation, Journal Year: 2023, Volume and Issue: 83, P. 101377 - 101377

Published: Aug. 14, 2023

Language: Английский

---

Effective and Imperceptible Adversarial Textual Attack via Multi-objectivization

ACM Transactions on Evolutionary Learning and Optimization, Journal Year: 2024, Volume and Issue: 4(3), P. 1 - 23

Published: July 23, 2024

The field of adversarial textual attack has significantly grown over the past few years, where commonly considered objective is to craft examples (AEs) that can successfully fool target model. However, imperceptibility attacks, which also essential for practical attackers, often left out by previous studies. In consequence, crafted AEs tend have obvious structural and semantic differences from original human-written text, making them easily perceptible. this work, we advocate leveraging multi-objectivization address such an issue. Specifically, reformulate problem crafting as a multi-objective optimization problem, auxiliary objective. Then, propose simple yet effective evolutionary algorithm, dubbed HydraText, solve problem. HydraText be effectively applied both score-based decision-based settings. Exhaustive experiments involving 44,237 instances demonstrate consistently achieves competitive success rates better than recently proposed approaches. A human evaluation study shows are more indistinguishable text. Finally, these exhibit good transferability bring notable robustness improvement model training.

Language: Английский

---

Related Keyframe Optimization Gaussian–Simultaneous Localization and Mapping: A 3D Gaussian Splatting-Based Simultaneous Localization and Mapping with Related Keyframe Optimization

Applied Sciences, Journal Year: 2025, Volume and Issue: 15(3), P. 1320 - 1320

Published: Jan. 27, 2025

Simultaneous localization and mapping (SLAM) is the basis for intelligent robots to explore world. As a promising method 3D reconstruction, Gaussian splatting (3DGS) integrated with SLAM systems has shown significant potential. However, due environmental uncertainties, errors in tracking process Gaussians can negatively impact systems. This paper introduces novel dense RGB-D system based on 3DGS that refines through sub-Gaussians camera coordinate system. Additionally, we propose an algorithm select keyframes closely related current frame, optimizing scene map pose of keyframe. approach effectively enhances both performance. Experiments high-quality synthetic scenes (Replica dataset) low-quality real-world (TUM-RGBD ScanNet datasets) demonstrate our achieves competitive performance mapping.

Language: Английский

---

On the Energy Consumption of ATL Transformations

Software Practice and Experience, Journal Year: 2025, Volume and Issue: unknown

Published: Feb. 14, 2025

Abstract Background Model transformations play a crucial role in Model‐Driven Engineering (MDE), with the ATLAS Transformation Language (ATL) being powerful technology for developing model‐to‐model transformations. Methods This paper presents comprehensive investigation into energy consumption of ATL transformations, aiming to identify possible correlations among transformation rules, model size, and metamodel structural characteristics. We conducted experiments on 52 analyzing power usage extending our inquiry understand impact mutations both models Results The experimental findings reveal relationships between utilization characteristics metamodels. Furthermore, we establish connection consumption, complexity processes. Conclusion insights gained from this research lay groundwork devising future energy‐efficient strategies while

Language: Английский

---

An efficient encoder-decoder network for the capacitated vehicle routing problem

Expert Systems with Applications, Journal Year: 2025, Volume and Issue: unknown, P. 127311 - 127311

Published: March 1, 2025

Language: Английский

---

The Pointer Network for Reward Maximisation in Multi-Target Space Mission Sequence Selection

Advances in Space Research, Journal Year: 2025, Volume and Issue: unknown

Published: April 1, 2025

Language: Английский

---