Edge Computing Optimization Using Mathematical Modeling, Deep Learning Models, and Evolutionary Algorithms

Ponnusamy, Vijayakumar; Rajalingam, Prithiviraj; Rȧjeswari, S.

doi:10.1002/9781119785521.ch2

Cited by 1 publication

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The forthcoming perspective of implementing the reliable high-speed 5G/6G cellular network standards provides permanent access to unbounded computational resources regardless of their possession and geographic location. This tendency appears in the areas of cloud [1,2], and edge [3,4] computing, computing power networks and synonims [5][6][7][8][9], and other systems of distributive computations. The providers have to develop the corresponding services under the conditions of the high creation, ownership, and operating costs, -the high user requirements stated in the Service Level Agreements (SLA), -the high requirements for the reaction time and efficiency of the computing resource management, -the high competition between the rival providers.…”

Section: Introductionmentioning

confidence: 99%

Stochastic Time Complexity Surfaces of Computing Node

Borisov,

Ivanov

2023

Mathematics

View full text Add to dashboard Cite

The paper is devoted to the formal description of the running time of the user task on some virtual nodes in the computing network. Based on the probability theory framework, this time represents a random value with a finite mean and variance. For any class of user task, these moments are the functions of the node resources, task numerical characteristics, and the parameters of the current node state. These functions of the vector arguments can be treated as some surfaces in the multidimensional Euclidean spaces, so the proposed models are called the stochastic time complexity surfaces. The paper also presents a class of functions suitable for the description of both the mean and variance. They contain unknown parameters which should be estimated. The article includes the statement of the parameter identification problem given the statistical results of the node stress testing, recommendations concerning the test planning, and preprocessing of the raw experiment data. To illustrate the performance of the proposed model, the authors design it for an actual database application—the prototype of the passengers’ personal data anonymization system. Its application functions are classified into two user task classes: the data anonymization procedures and fulfillment of the statistical queries. The authors identify the stochastic time complexity surfaces for both task types. The additional testing experiments confirm the high performance of the suggested model and its applicability to the solution of the practical providers’ problems.

show abstract