Jiaxin Wu

Jiaxin Wu

Auto-generated profile, based on publicly available information

Jiaxin Wu

School of Computer Science and Technology, Guangdong University of Technology, Guangzhou, China

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 December 2017
Verified
Conference Paper
Closed access
JWJiaxin Wu
Jiaxin Wu
HWHanpei Wei
Hanpei Wei
YZYi Zhang
Yi Zhang
HWHaifeng Wei
Haifeng Wei
The paper analyze the basic principle of model reference adaptive system(MRAS) algorithm. In this paper, rotor position and speed estimation algorithm of permanent magnet synchronous motor was deduced based on MRAS, and it is applied to sensorless vector control system of permanent magnet synchronous motor. Sensorless vector control system based on MRAS is built under the Matlab/Simulink environment. The simulation experiment and result analysis for the situation of motor's starting, speed mutation and load mutation are done. Simulation results prove that the sensorless vector control system of permanent magnet synchronous motor based on the algorithm of MRAS takes good control performance, and the algorithm can accurately estimate the rotor position and speed of the motor.

1 September 2021
Verified
Conference Paper
Closed access
SHSuna He
Suna He
JWJigang Wu
Jigang Wu
BWBing Wei
Bing Wei
JWJiaxin Wu
Jiaxin Wu
Tree-shaped task graphs become a paradigm to be utilized in distributed platform for various computational domains, such as the electronic structure calculations and the factorization of sparse matrices. However, the scheduling of the tree-shaped task graph has been rarely studied for the more realistic heterogeneous multiprocessor platform (HEMP). This paper proposes an efficient algorithm named Partition-Allocation (PA) for parallel computing on HEMP with limited memory. Algorithm PA consists of two stages: partitioning and allocation. In the partitioning stage, a task tree is split into several subtrees. In the allocation stage, these subtrees are assigned to different processors for execution. Our algorithm PA can reduce makespan by prioritizing subtrees on the critical path, both in the partitioning and in the allocation. Based on randomly generated trees and real-world dataset, experimental results show that the proposed PA is significantly better than the latest work in terms of average makespan. The proposed algorithm can successfully reduce the average makespan by up to 67.01% on real-world dataset, and 52.35% on randomly generated trees.

14 March 2025 in IEEE Transactions on Network Science and Engineering
Verified
Journal Article
Closed access
JHJiale Huang
Jiale Huang
JWJigang Wu
Jigang Wu
YWYalan Wu
Yalan Wu
JWJiaxin Wu
Jiaxin Wu
Extensive research on edge inference has devoted in optimizing service performance for users. However, recent studies have overlooked the desired utility of application service provider (ASP), which is crucial for achieving long-term service provisioning. Besides, efficient request offloading and resource allocation are essential for optimizing long-term utility of ASP in dynamic networks with time-coupled resources. To address these issues, this paper formulates a long-term utility optimization problem in collaborative edge inference system. The objective is to maximize the long-term average utility of ASP, by jointly optimizing request offloading and resource allocation, under the time-coupled resource constraints. To solve the problem, a Lyapunov based online algorithm is proposed to decompose it into a series of one-slot deterministic problems by decoupling the time-coupled resource constraints. Only the current network states are required for one-slot problem. Then, the one-slot problem is converted into a master request offloading problem with an inner resource allocation problem. A distributed algorithm is proposed to derive the optimal decision to inner problem, while a coalition based algorithm is proposed to seek the stable solution to master problem. Experimental results show that, the proposed algorithm outperforms baseline algorithms for most cases, in terms of long-term average utility of ASP.

1 December 2019 in Physical Communication
Verified
Journal Article
Closed access
JWJiaxin Wu
Jiaxin Wu
JWJigang Wu
Jigang Wu
LCLong CHEN
Long CHEN
YWYalan Wu
Yalan Wu

22 June 2023 in IEEE Transactions on Network and Service Management
Verified
Journal Article
Closed access
JWJiaxin Wu
Jiaxin Wu
JWJigang Wu
Jigang Wu
LCLong CHEN
Long CHEN
YSYifei Sun
Yifei Sun
YWYalan Wu
Yalan Wu
The intrinsic nature of non-independent and identically distributed datasets on heterogeneous devices slows down the distributed model training process and reduces the training accuracy. To settle this problem, we propose a dataset reconstruction scheme to transform the data distribution of training device’s dataset into independent and identically distributed dataset via data exchange among trusted devices. For energy efficiency, we further consider power control for the devices. We then formulate an optimization problem, which is a mixed integer non-linear programming problem, to minimize the total energy consumption for each round of distributed training. Due to the NP-hardness and coupling property of the optimization problem, we decompose it into two subproblems for dataset reconstruction and power control, respectively. An approximation algorithm is designed to obtain a near-optimal auxiliary devices set for dataset reconstruction with minimum energy consumption, while meeting the variance constraint of the optimization problem. We prove that approximation algorithm has a worst-case approximation ratio of 1+ln|Ωi(t)|, where |Ωi(t)| is the required data samples for dataset reconstruction of each training device. For power control, we design a dynamic programming algorithm to further reduce the energy consumption. For comparison, we propose three benchmark schemes that adopt either one of the algorithms or neither. We also customize three baseline algorithms based on the state-of-the-arts to compare with our proposed algorithm. Numerical results show that, our proposed algorithm outperforms three benchmarks on the average energy consumption for one round for different cases. When varying the labels that each device owns, our proposed algorithm outperforms the other three baseline algorithms on training accuracy. Besides, when setting a target accuracy, our proposed algorithm always has the lowest energy consumption.

29 May 2026 in IEEE Transactions on Mobile Computing
Verified
Journal Article
Closed access
JHJiale Huang
Jiale Huang
YWYalan Wu
Yalan Wu
JWJigang Wu
Jigang Wu
JWJiaxin Wu
Jiaxin Wu
SLSiew Kei Lam
Siew Kei Lam
Recent endeavors on edge inference focus on system wide service performance optimization. However, they overlook fairness guarantee in request completion rates across users, which plays a pivotal role in enhancing user satisfaction. This paper fills this gap by formulating a long-term max-min user fairness problem in collaborative edge inference systems. The goal is to maximize the minimum request completion rate among users, by optimizing request offloading and resource allocation, under constraints of quality-of-service, resources, and time-coupled energy budgets. An online fairness guarantee algorithm is proposed to solve the formulated problem. The online algorithm transforms the problem into an equivalent maximization problem, and then decomposes it into a series of one-slot problems by decoupling time-coupled constraints. The online algorithm solves one-slot problem in dual domain, where the dual function is decomposed into request offloading problem and resource allocation problem. Agreedy strategy is proposed to solve request offloading problem in parallel, while convex optimization techniques are utilized to solve resource allocation problem in parallel. The asymptotic optimality of the online algorithm is proved. Experimental results show that the proposed algorithm outperforms baseline algorithms for most cases, in terms of minimum request completion rate, while consistently achieving near-perfect fairness.