Cell Multi-Agent Coordination for Shared Control Elements
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional AI-based solutions for optimizing cell parameters in cellular networks fail to account for the impact of parameter changes in one cell on other cells sharing a common control element, leading to inconsistent and inefficient optimization.
Innovation Solution
A method and apparatus that consolidate intermediate control decisions from multiple cell agents into a single consolidated control decision using a coordination agent, which is independent of the individual cell agents, to control shared network elements such as antenna tilts across multiple cells.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional AI-based solutions optimize cell parameters independently per cell, then each cell can be optimized individually with simple agent functions, but the parameter changes in one cell negatively impact other cells sharing the same control element
Solution Approach 1:
The patent introduces a coordination agent as an intermediary between individual cell agents and the control element. This coordination agent receives intermediate control decisions from multiple cell agents, consolidates them into a single consolidated control decision, and sends it to the control element. This mediator resolves the conflict between independent optimization and coordinated control by translating multiple independent decisions into one unified decision that accounts for shared control elements.
2Reliability
If a single distributed RL agent is used to optimize cell parameters, then coordination among cells is improved, but processing capabilities are overwhelmed and efficiency decreases
Solution Approach 1:
The patent segments the optimization system into multiple independent cell agents, each responsible for a single cell, and a separate coordination agent. Each cell agent independently processes its own cell's data and generates intermediate control decisions without requiring complex processing of entire network data. This segmentation distributes the computational load, avoiding the processing bottleneck of a single centralized agent while maintaining coordination through the consolidation mechanism.
3Productivity
If conventional solutions consistently select the most appropriate value for one cell sharing a control element, then optimization for that specific cell is maximized, but the impact on other sharing cells is disregarded leading to suboptimal overall performance
Solution Approach 1:
The patent merges the intermediate control decisions from multiple cell agents into a single consolidated control decision. The coordination agent combines the decisions by evaluating the shared control element's constraints and selecting a unified decision that satisfies all cells sharing that element. This merging process ensures that the optimization considers all affected cells simultaneously, achieving both speed (through automated consolidation) and reliability (through coordinated optimization).
Data Source
AI summary
A method and apparatus extend existing cell parameter optimization techniques so that they can be used in cases where different cells are controlled by the same control element. Particularly, a coordination agent (46) receives an intermediate control decision for controlling a control element (48) shared by the plurality of cells (14). The coordination agent consolidates each of the received intermediate decisions into a single consolidated control decision, and sends the consolidated control decision to the shared control element. The control element then controls a network element (26) in each of the cells using the same consolidated control decision.