Cross-Entropy Method for Electromagnetic Optimization With Constraints and Mixed Variables

Kovaleva, Maria; Bulger, David; Zeb, Basit Ali

doi:10.1109/tap.2017.2740974

Cited by 31 publications

(12 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For probability learning, the probability distribution function p(x) is usually introduced with an indicator u, e.g., p(x, u) can be a Gaussian distribution and u contains its mean and variance [11]. Denoting that L equals to N × (M + 1), the indicator u is a vector of L dimensions, defined as…”

Section: B the Asce-based Offload Learningmentioning

confidence: 99%

A Novel Cross Entropy Approach for Offloading Learning in Mobile Edge Computing

Zhu

Fan

et al. 2020

IEEE Wireless Commun. Lett.

View full text Add to dashboard Cite

In this paper, we propose a novel offloading learning approach to compromise energy consumption and latency in a multi-tier network with mobile edge computing. In order to solve this integer programming problem, instead of using conventional optimization tools, we apply a cross entropy approach with iterative learning of the probability of elite solution samples. Compared to existing methods, the proposed one in this network permits a parallel computing architecture and is verified to be computationally very efficient. Specifically, it achieves performance close to the optimal and performs well with different choices of the values of hyperparameters in the proposed learning approach.

show abstract

Section: B the Asce-based Offload Learningmentioning

confidence: 99%

A Novel Cross Entropy Approach for Offloading Learning in Mobile Edge Computing

Zhu

Fan

et al. 2020

IEEE Wireless Commun. Lett.

View full text Add to dashboard Cite

show abstract

“…In particular, from the constant directivity contour curves (corrected Carrel's graph, [1]) and considering antenna length equal to 0.619m, the optimum value of σ is derived equal to 0.1648 and the respective value of τ equal to 0.8891, while the derived antenna has to be composed of 10 dipoles. The largest dipole ( 1 m = ) is considered to be in resonant condition at the lowest frequency ( min 470MHz f = ) of the passband, and therefore its length must be equal to λmax/2, as shown in (8). Also, we decided to set the radius of the shortest dipole ( 10) = m equal to the radius (2 mm) of the dipoles of the optimized LPDA, and therefore the other (larger) dipoles of the LPDA will have larger radii, resulting thus in an antenna that can be fabricated in practice.…”

Section: Definition Of Limits Of Optimization Variablesmentioning

confidence: 99%

Optimal LTE-Protected LPDA Design for DVB-T Reception Using Particle Swarm Optimization With Velocity Mutation

Zaharis

Gravas

Lazaridis

et al. 2018

IEEE Trans. Antennas Propagat.

View full text Add to dashboard Cite

A near-optimal design of a log-periodic dipole array (LPDA), suitable for DVB-T reception (470-790 MHz), is presented. The LPDA is required to provide low standing wave ratio as well as high-gain radiation pattern with sufficient gain flatness over the entire passband, and concurrently achieve low gain for frequencies above 800 MHz to reject LTE800 signals and thus improve the reception quality in the DVB-T band. All the above requirements are better satisfied by applying a novel particle swarm optimization (PSO) variant, called PSO with velocity mutation (PSOvm). PSOvm induces mutation on the velocities of those particles, which are unable to improve their fitness. As shown in this paper, PSOvm comes closer to the above requirements compared to four well-known optimization methods and outperforms the traditional LPDA design method proposed by Carrel. The LPDA geometry chosen for optimization is not the conventional one and therefore the dipoles are not considered to be included inside a specified angle as proposed by Carrel. Thus, the dipole lengths and distances as well as the boom dimensions are independently optimized. The PSOvm-based LPDA sufficiently meets all the above requirements and thus is suitable for DVB-T reception without the use of an external LTE-band rejection filter.

show abstract

“…where x [s] is the association corresponding to the sth item in the resorted sequence obtained from step 5 in Algorithm 1. By following the procedure in each iteration, as inspired in [6], the CE approach can produce a sequence of sampling distributions that are increasingly concentrated around the optimal design.…”

Section: A Problem Formulation Of Association Learningmentioning

confidence: 99%

“…The CE approach was firstly introduced in 1997 [5] and developed in machine learning. The advantage of the CE approach lies in its adaptive update procedure [6], which makes it be inherently capable of solving combinatorial optimization problems in a much simpler way than typical relaxation techniques. To the best of our knowledge, this is the first time that CE method is used to solve the constrained user association problem.…”

Section: Introductionmentioning

confidence: 99%

Learning Oriented Cross-Entropy Approach to User Association in Load-Balanced HetNet

Huang

Xie

et al. 2018

IEEE Wireless Commun. Lett.

View full text Add to dashboard Cite

This letter considers optimizing user association in a heterogeneous network via utility maximization, which is a combinatorial optimization problem due to integer constraints. Different from existing solutions based on convex optimization, we alternatively propose a cross-entropy (CE)-based algorithm inspired by a sampling approach developed in machine learning. Adopting a probabilistic model, we first reformulate the original problem as a CE minimization problem which aims to learn the probability distribution of variables in the optimal association. An efficient solution by stochastic sampling is introduced to solve the learning problem. The integer constraint is directly handled by the proposed algorithm, which is robust to network deployment and algorithm parameter choices. Simulations verify that the proposed CE approach achieves near-optimal performance quite efficiently.

show abstract

Cross-Entropy Method for Electromagnetic Optimization With Constraints and Mixed Variables

Cited by 31 publications

References 38 publications

A Novel Cross Entropy Approach for Offloading Learning in Mobile Edge Computing

A Novel Cross Entropy Approach for Offloading Learning in Mobile Edge Computing

Optimal LTE-Protected LPDA Design for DVB-T Reception Using Particle Swarm Optimization With Velocity Mutation

Learning Oriented Cross-Entropy Approach to User Association in Load-Balanced HetNet

Contact Info

Product

Resources

About