This paper studies a federated edge learning system,in which an edge server coordinates a set of edge devices to train a shared machine learning(ML)model based on their locally distributed data samples.During the dist...This paper studies a federated edge learning system,in which an edge server coordinates a set of edge devices to train a shared machine learning(ML)model based on their locally distributed data samples.During the distributed training,we exploit the joint communication and computation design for improving the system energy efficiency,in which both the communication resource allocation for global ML-parameters aggregation and the computation resource allocation for locally updating ML-parameters are jointly optimized.In particular,we consider two transmission protocols for edge devices to upload ML-parameters to edge server,based on the non-orthogonal multiple access(NOMA)and time division multiple access(TDMA),respectively.Under both protocols,we minimize the total energy consumption at all edge devices over a particular finite training duration subject to a given training accuracy,by jointly optimizing the transmission power and rates at edge devices for uploading ML-parameters and their central processing unit(CPU)frequencies for local update.We propose efficient algorithms to solve the formulated energy minimization problems by using the techniques from convex optimization.Numerical results show that as compared to other benchmark schemes,our proposed joint communication and computation design significantly can improve the energy efficiency of the federated edge learning system,by properly balancing the energy tradeoff between communication and computation.展开更多
For the purpose of solving the engineering constrained discrete optimization problem, a novel discrete particle swarm optimization（DPSO） is proposed. The proposed novel DPSO is based on the idea of normal particle s...For the purpose of solving the engineering constrained discrete optimization problem, a novel discrete particle swarm optimization（DPSO） is proposed. The proposed novel DPSO is based on the idea of normal particle swarm optimization（PSO）, but deals with the variables as discrete type, the discrete optimum solution is found through updating the location of discrete variable. To avoid long calculation time and improve the efficiency of algorithm, scheme of constraint level and huge value penalty are proposed to deal with the constraints, the stratagem of reproducing the new particles and best keeping model of particle are employed to increase the diversity of particles. The validity of the proposed DPSO is examined by benchmark numerical examples, the results show that the novel DPSO has great advantages over current algorithm. The optimum designs of the 100-1 500 mm bellows under 0.25 MPa are fulfilled by DPSO. Comparing the optimization results with the bellows in-service, optimization results by discrete penalty particle swarm optimization（DPPSO） and theory solution, the comparison result shows that the global discrete optima of bellows are obtained by proposed DPSO, and confirms that the proposed novel DPSO and schemes can be used to solve the engineering constrained discrete problem successfully.展开更多
Extensibility and attainability of topology optimization are discussed by investigating a variety of simultaneous topology opti-mization methods extended from the standard formulation.First,the state of the art is hig...Extensibility and attainability of topology optimization are discussed by investigating a variety of simultaneous topology opti-mization methods extended from the standard formulation.First,the state of the art is highlighted through systematic classification of developed methods,such as simultaneous topology optimizations of microstructure and macrostructure,structure and supports,structure and design-dependent loads,structure and locations of involved components.Second,some recent results about simultaneous topology optimization of structure and applied loads are presented.It is shown that the simultaneous topology optimization is an integrated methodology that extends the concept of standard topology optimization in the sense of systematic design.The presence of more than one kind of design variable of different nature makes the optimization problem complex but enlarges the design space to attain the optimization.展开更多
基金the National Key R&D Program of China under Grant 2018YFB1800800Guangdong Province Key Area R&D Program under Grant 2018B030338001the Natural Science Foundation of China under Grant U2001208。
文摘This paper studies a federated edge learning system,in which an edge server coordinates a set of edge devices to train a shared machine learning(ML)model based on their locally distributed data samples.During the distributed training,we exploit the joint communication and computation design for improving the system energy efficiency,in which both the communication resource allocation for global ML-parameters aggregation and the computation resource allocation for locally updating ML-parameters are jointly optimized.In particular,we consider two transmission protocols for edge devices to upload ML-parameters to edge server,based on the non-orthogonal multiple access(NOMA)and time division multiple access(TDMA),respectively.Under both protocols,we minimize the total energy consumption at all edge devices over a particular finite training duration subject to a given training accuracy,by jointly optimizing the transmission power and rates at edge devices for uploading ML-parameters and their central processing unit(CPU)frequencies for local update.We propose efficient algorithms to solve the formulated energy minimization problems by using the techniques from convex optimization.Numerical results show that as compared to other benchmark schemes,our proposed joint communication and computation design significantly can improve the energy efficiency of the federated edge learning system,by properly balancing the energy tradeoff between communication and computation.
基金supported by National Hi-tech Research and Development Program of China (Grant No. 2006aa042439)
文摘For the purpose of solving the engineering constrained discrete optimization problem, a novel discrete particle swarm optimization（DPSO） is proposed. The proposed novel DPSO is based on the idea of normal particle swarm optimization（PSO）, but deals with the variables as discrete type, the discrete optimum solution is found through updating the location of discrete variable. To avoid long calculation time and improve the efficiency of algorithm, scheme of constraint level and huge value penalty are proposed to deal with the constraints, the stratagem of reproducing the new particles and best keeping model of particle are employed to increase the diversity of particles. The validity of the proposed DPSO is examined by benchmark numerical examples, the results show that the novel DPSO has great advantages over current algorithm. The optimum designs of the 100-1 500 mm bellows under 0.25 MPa are fulfilled by DPSO. Comparing the optimization results with the bellows in-service, optimization results by discrete penalty particle swarm optimization（DPPSO） and theory solution, the comparison result shows that the global discrete optima of bellows are obtained by proposed DPSO, and confirms that the proposed novel DPSO and schemes can be used to solve the engineering constrained discrete problem successfully.
基金supported by the National Natural Science Foundation of China(Grant Nos.51275424,51221001)the National Basic Research Program of China("973"Project)(Grant No.2011CB610304)+1 种基金the Research Fund for the Doctoral Program of Higher Education of China(Grant No.20126102130003)the NWPU Foundation for Fundamental Research(Grant No.NPU-FFR-201001)
文摘Extensibility and attainability of topology optimization are discussed by investigating a variety of simultaneous topology opti-mization methods extended from the standard formulation.First,the state of the art is highlighted through systematic classification of developed methods,such as simultaneous topology optimizations of microstructure and macrostructure,structure and supports,structure and design-dependent loads,structure and locations of involved components.Second,some recent results about simultaneous topology optimization of structure and applied loads are presented.It is shown that the simultaneous topology optimization is an integrated methodology that extends the concept of standard topology optimization in the sense of systematic design.The presence of more than one kind of design variable of different nature makes the optimization problem complex but enlarges the design space to attain the optimization.
