Self-Adaptation Resource Allocation for Continuous Offloading Tasks in Pervasive Computing

Area of Science:

• Computer Science
• Artificial Intelligence
• Mobile Computing

Background:

• Technological advancements increase data volume and user expectations for sophisticated device features.
• High-definition displays and sensor usage rapidly deplete device battery power.
• Low latency requirements can exacerbate power consumption in smart devices.

Purpose of the Study:

• To investigate the effectiveness of cloud computing and metaheuristic algorithms in optimizing task offloading for mobile devices.
• To reduce energy consumption and improve the efficiency of smart devices.
• To compare the performance of different metaheuristic algorithms for task offloading.

Main Methods:

• Offloading computational tasks from smart devices to cloud servers.
• Simulating tasks in a virtual environment on cloud servers to analyze resource parameters.
• Utilizing metaheuristic algorithms, including First-Come, First-Served (FCFS), Ant Colony Optimization (ACO), and Particle Swarm Optimization (PSO), for task offloading decisions.
• Comparing battery consumption and makespan time across different algorithms.

Main Results:

• Particle Swarm Optimization (PSO) demonstrated lower battery consumption and makespan time compared to FCFS and ACO.
• Offloading device resources to the cloud effectively reduces overall energy consumption.
• PSO enhances device battery life and system efficiency.

Conclusions:

• Cloud computing, combined with optimized metaheuristic algorithms like PSO, offers a viable solution for managing computational load and conserving energy on mobile devices.
• PSO is a superior algorithm for task offloading, leading to significant improvements in device battery life and system performance.
• The study highlights the potential of intelligent offloading strategies to meet the demands of modern, data-intensive applications.