.png){kind=icon}
.png){kind=icon}
Return
Distributed Optimization Algorithm Design and Analysis on Cooperation-Competition Network Based on PID Control
.png){kind=icon}
PRE
AI
DOI：10.1016/j.jfranklin.2026.108483.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We propose a novel distributed optimization algorithm on the cooperation-competition network based on proportional-integral-derivative (PID) control to enhance the convergence rate. • We demonstrate theoretically that the algorithm achieves global exponential convergence to an optimal solution when the local objective functions are smooth and strongly convex and provide guidelines for selecting appropriate parameter values (e.g., kp, ki, kd). • The simulation results support the theoretical findings and provide empirical evidence for the performance and efficiency of the proposed approaches. Furthermore, comparative experiments demonstrate the flexibility of our algorithm. • We explore that the D-PID-CCN has great potential for non-convex distributed optimization over the cooperation-competition network. Simulation results imply that the term ki1t∫0t∇fi(xi(s))ds may facilitate the objective function in escaping from local minimum points and reaching the global minimum point.
Keywords:
Distributed Optimization
Cooperation-Competition Network
PID Control
Global Exponential Convergence
Non-Convex Optimization
Journal
J
IF:
4.2
Papers:
822
Citations:
0