Abstract:This paper is concerned with solving a large category of convex optimization problems using a group of agents, each only being accessible to its individual convex cost function. The optimization problems are modeled as minimizing the sum of all the agents' cost functions. The communication process between agents is described by a sequence of time-varying yet balanced directed graphs which are assumed to be uniformly strongly connected. Taking into account the fact that the communication channel bandwidth is limited, for each agent we introduce a vector-valued quantizer with finite quantization levels to preprocess the information to be exchanged. We exploit an event-triggered broadcasting technique to guide information exchange, further reducing the communication cost of the network. By jointly designing the dynamic event-triggered encoding-decoding schemes and the event-triggered sampling rules (to analytically determine the sampling time instant sequence for each agent), a distributed subgradient descent algorithm with constrained information exchange is proposed. By selecting the appropriate quantization levels, all the agents' states asymptotically converge to a consensus value which is also the optimal solution to the optimization problem, without committing saturation of all the quantizers. We find that one bit of information exchange across each connected channel can guarantee that the optimiztion problem can be exactly solved. Theoretical analysis shows that the event-triggered subgradient descent algorithm with constrained data rate of networks converges at the rate of O(ln t/root t). We supply a numerical simulation experiment to demonstrate the effectiveness of the proposed algorithm and to validate the correctness of theoretical results.

Quantized Subgradient Algorithm with Limited Bandwidth Communications for Solving Distributed Optimization over General Directed Multi-Agent Networks.

Quantized Subgradient Algorithm and Data-Rate Analysis for Distributed Optimization

Distributed Subgradient Method for Multi‐agent Optimization with Quantized Communication

Distributed subgradient method for multi-agent optimization with quantized communication: J. LIET AL

Distributed Constrained Optimization With Delayed Subgradient Information Over Time-Varying Network Under Adaptive Quantization

Distributed Quantized Gradient-Free Algorithm for Multi-Agent Convex Optimization

Computation-Efficient Distributed Algorithm for Convex Optimization over Time-Varying Networks with Limited Bandwidth Communication

Distributed Consensus Optimization in Multiagent Networks with Time-Varying Directed Topologies and Quantized Communication.

Distributed dual averaging method for multi-agent optimization with quantized communication.

Distributed Optimization with Quantized Gradient Descent

Distributed Economic Dispatch Algorithm with Quantized Communication Mechanism

Quantized Gradient-Descent Algorithm for Distributed Resource Allocation

Convergence Theory of Generalized Distributed Subgradient Method with Random Quantization

Distributed Optimization via Gradient Descent with Event-Triggered Zooming over Quantized Communication

Distributed Online Optimization Under Dynamic Adaptive Quantization

Effects of Probabilistic Quantization on Distributed Multi-agent Optimization Problem

Distributed Quantized Optimization Design of Continuous-Time Multiagent Systems over Switching Graphs.

Distributed Subgradient Method with Random Quantization and Flexible Weights: Convergence Analysis.

A Distributed Algorithm with a Modified Quantization Radius under Limited Communication

Event-Triggered Communication and Data Rate Constraint for Distributed Optimization of Multiagent Systems.

Distributed Quantized Mirror Descent for Strongly Convex Optimization over Time-Varying Directed Graph