DISH: A Distributed Hybrid Primal-Dual Optimization Framework to Utilize System Heterogeneity

Keyphrases

• Optimization Framework 100%
• Primal-dual Optimization 100%
• Newton-type 100%
• System Heterogeneity 100%
• Gradient Type 66%
• Distributed Cooperative Optimization 66%
• Hybrid Method 33%
• Computational Cost 33%
• Algorithm Framework 33%
• Optimization Problem 33%
• Hessian 33%
• Network Topology 33%
• General Network 33%
• Computational Capability 33%
• Local Computation 33%
• Primal-dual 33%
• Distributed Method 33%
• Local Update 33%
• Provable Convergence 33%
• Computational Capacity 33%
• Convergence Guarantee 33%
• Linear Convergence 33%
• Approximate Newton Method 33%
• Multi-agent Networks 33%
• Q-linear Convergence 33%
• Strongly Convex Function 33%
• Exact Optimal Solution 33%
• Rate Guarantees 33%

Mathematics

• Computational Cost 100%
• Approximates 100%
• Numerical Analysis 100%
• Linear Convergence 100%
• Convergence Rate 100%
• Newton's Method 100%
• Convex Function 100%

Computer Science

• Optimization Framework 100%
• Primal-Dual 100%
• System Heterogeneity 100%
• Optimization Problem 50%
• Convex Function 50%
• Network Topology 50%
• multi agent 50%
• Computational Cost 50%
• Convergence Rate 50%
• Newton's Method 50%