Application of an improved Discrete Salp Swarm Algorithm to the wireless rechargeable sensor network problem.

Area of Science:

• Computer Science
• Artificial Intelligence
• Optimization Algorithms

Background:

• Swarm intelligence algorithms are crucial for solving complex optimization problems.
• Existing algorithms like the Ant Colony System (ACS) have limitations in efficiency and accuracy for certain tasks.
• The Traveling Salesperson Problem (TSP) and Wireless Rechargeable Sensor Network (WRSN) pathfinding are computationally intensive challenges.

Purpose of the Study:

• To develop an improved Discrete Salp Swarm Algorithm (DSSACS) by integrating the Ant Colony System (ACS).
• To enhance the initialization, discretization, and foraging simulation within the salp swarm algorithm.
• To evaluate the performance of DSSACS on TSP and WRSN path planning problems compared to existing methods.

Main Methods:

• Optimizing salp colony initialization using the Ant Colony System (ACS).
• Discretizing the salp swarm algorithm for combinatorial optimization.
• Simulating salp foraging behavior using crossover and mutation operators.
• Benchmarking DSSACS against other algorithms on various TSP datasets.
• Applying DSSACS to multi-path planning in Wireless Rechargeable Sensor Networks (WRSNs).

Main Results:

• DSSACS achieved significantly lower error rates on TSP datasets (0.78%-2.95%) compared to other algorithms (2.03%-6.43%).
• The proposed algorithm exhibits faster convergence speed, a robust positive feedback mechanism, and higher accuracy.
• For WRSN path planning, DSSACS selected paths approximately 20% shorter than those selected by ACS.
• DSSACS demonstrated clear advantages in multi-path planning for mobile charging vehicles (MCVs) within WRSNs.

Conclusions:

• The integration of ACS into the Discrete Salp Swarm Algorithm (DSSACS) yields a more effective optimization tool.
• DSSACS offers superior performance in terms of accuracy, convergence, and path optimization for TSP and WRSN applications.
• The enhanced algorithm provides significant time and cost savings compared to other swarm intelligence approaches in WRSN scenarios.