We tackle the problem of having multiple transmitters cooperating to be desynchronized using a distributed algorithm. Although this problem can also be found in surveillance, it has the most impact in achieving a fair...
详细信息
We tackle the problem of having multiple transmitters cooperating to be desynchronized using a distributed algorithm. Although this problem can also be found in surveillance, it has the most impact in achieving a fair access to a wireless shared communication medium at the medium access control layer in the context of wireless sensor networks. In this article, we first theoretically investigate the convergence rate of various optimization algorithms, giving closed-form expressions for the parameters achieving the best worst-case convergence rate. We then show that a recently proposed time-varying parameters Nesterov algorithm applied to this problem has worse performance assuming one can determine the number of sensors in the network. In order to remove such an assumption, the problem is seen as the solution of a linear equation corresponding to the first optimality condition. Both theoretically and in simulation, we show that using the Gauss-Seidel method improves the speed of convergence, although its performance degrades for large network sizes. In simulations, it is shown the behavior for various number of wireless devices, emphasizing how the algorithms actually perform in comparison with their worst-case theoretical rates for different network sizes.
Genetic algorithm has been widely used in route planning problems due to its strong global search ability by simulating material changes in nature, as well as biological activities and evolution processes. Predation s...
详细信息
This report is all about the invention called the K1 Fluu, a new method in fast learning neural networks. K1 Fluu gets gradient slopes easier and yields hidden weights higher. It applies higher order derivatives, weig...
详细信息
This paper addresses the shortcomings of the novel metaheuristic algorithm, Crested Porcupine Optimizer (CPO), which tends to fall into local optima during optimization, exhibits slow convergence in the later stages o...
详细信息
The exploration of advanced algorithms for solving Sudoku puzzles reveals unique insights into their performance and effectiveness. The study analyzes and compares effectiveness in solving Sudoku puzzles with the clas...
详细信息
This paper proposes an index-based octree neighborhood particle search algorithm to optimize the efficiency of neighborhood particle search in Smoothed Particle Hydrodynamics (SPH) simulations. Based on the octree sea...
详细信息
As the number and complexity of constraints in constrained multi-objective optimization problems (CMOPs) increase, the performance of existing constrained multi-objective evolutionary algorithms (CMOEAs) declines sign...
详细信息
One of the most important problems of the Differ-ential Evolution algorithm is the adaptation of scaling factor, due to high sensitivity to this parameter. In this paper the L-SRTDE algorithm is proposed, where the sc...
详细信息
STOchastic Recursive Momentum (STORM)based algorithms have been widely developed to solve one to K-level (K ≥ 3) stochastic optimization problems. Specifically, they use estimators to mitigate the biased gradient iss...
详细信息
STOchastic Recursive Momentum (STORM)based algorithms have been widely developed to solve one to K-level (K ≥ 3) stochastic optimization problems. Specifically, they use estimators to mitigate the biased gradient issue and achieve near-optimal convergence results. However, there is relatively little work on understanding their generalization performance, particularly evident during the transition from one to K-level optimization contexts. This paper provides a comprehensive generalization analysis of three representative STORM-based algorithms: STORM, COVER, and SVMR, for one, two, and K-level stochastic optimizations under both convex and strongly convex settings based on algorithmic stability. Firstly, we define stability for K-level optimizations and link it to generalization. Then, we detail the stability results for three prominent STORM-based algorithms. Finally, we derive their excess risk bounds by balancing stability results with optimization errors. Our theoretical results provide strong evidence to complete STORM-based algorithms: (1) Each estimator may decrease their stability due to variance with its estimation target. (2) Every additional level might escalate the generalization error, influenced by the stability and the variance between its cumulative stochastic gradient and the true gradient. (3) Increasing the batch size for the initial computation of estimators presents a favorable trade-off, enhancing the generalization performance. Copyright 2024 by the author(s)
With the transformation of economic development mode, the acceleration of urbanization, the optimization and upgrading of value energy structure, the structure of distribution network is becoming more and more complex...
详细信息
暂无评论