Online Convex Optimization Algorithm for Multi-Cell MIMO Precoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online convex optimization (OCO) algorithms for downlink multi-cell multiple input multiple output (MIMO) wireless network virtualization are limited by strict per-time-slot settings, which do not accommodate decisions and system updates that span multiple time slots, and fail to handle delayed, out-of-order, and missing gradient or sub-gradient feedbacks, leading to suboptimal performance in dynamic wireless environments.
Innovation Solution
An OCO algorithm with periodic updates is developed, allowing decisions to be made at the beginning of each update period lasting multiple time slots, with gradient or sub-gradient feedbacks delayed, received out of order, and partly missing, while minimizing regret and long-term constraint violations by utilizing past feedback information and enabling simultaneous sharing of antennas and spectrum resources among service providers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If decisions are made at each time slot with full information, then optimal performance is achieved, but computational complexity and system overhead increase significantly
Solution Approach 1:
The patent segments the continuous decision-making process into discrete update periods. Decisions are made periodically rather than continuously at each time slot, reducing computational overhead while maintaining acceptable performance. The system divides time into update intervals where decisions remain fixed, and information is refreshed only at period boundaries.
Solution Approach 2:
The patent implements periodic update mechanism where system parameters and decisions are updated at regular intervals rather than continuously. This periodic action reduces the frequency of computational operations and information exchanges, lowering system complexity while maintaining decision quality through sufficiently frequent updates.
2Device complexity
If decisions are made periodically over multiple time slots, then computational complexity is reduced, but system adaptability to changing conditions deteriorates
Solution Approach 1:
The patent introduces dynamic adjustment of the update period length based on system conditions. When channel conditions change rapidly, the update period is shortened to improve adaptability. When conditions are stable, the update period is extended to reduce complexity. This dynamic tuning allows the system to adapt its own operation to match environmental changes.
Solution Approach 2:
The patent implements feedback mechanisms where system performance and channel condition changes are monitored, and this feedback is used to adjust the update frequency and decision parameters. The feedback loop enables the periodic system to respond to changing conditions by modifying its update rhythm, thereby maintaining adaptability despite periodic operation.
3Measurement precision
If complete gradient feedback is received at each time slot, then optimization accuracy is maximized, but communication overhead and information loss increase
Solution Approach 1:
The patent merges gradient information from multiple time slots into a single aggregated feedback signal that is processed at update periods. Instead of handling complete gradient feedback at each time slot, the system combines gradients over the update period, reducing the total amount of information transmitted and processed while maintaining optimization accuracy through cumulative information.
Solution Approach 2:
The patent performs preliminary accumulation of gradient information during the update period before processing. Gradients are collected and aggregated in advance, then processed together at the update moment. This preliminary action reduces the immediate communication overhead and information loss by batching transmissions rather than sending complete feedback continuously.
4Reliability
If complete gradient feedback is received, then regret minimization is improved, but handling delayed and out-of-order feedback increases system complexity
Solution Approach 1:
The patent implements a self-synchronizing feedback mechanism where the periodic update structure automatically handles out-of-order and delayed feedback. The system uses the periodic decision update points as synchronization anchors, and feedback is associated with the appropriate update period based on timing information. This self-service approach eliminates the need for complex reordering protocols, as the periodic structure naturally accommodates feedback arrival variations.
Data Source
AI summary
A method and network node for online coordinated multi-cell precoding are provided. According to one aspect, a method includes receiving from each of the plurality of service providers a virtual precoder matrix determined by the corresponding service provider. The method also includes determining a precoder matrix by minimizing a precoding deviation from a virtualization demand of the network subject to at least one power constraint, the virtualization demand being based at least in part on a product of a channel state matrix and a virtual precoder matrix. The method further includes applying the determined precoder matrix to signals applied to a plurality of antennas to achieve a sum of throughputs for the plurality of service providers that is greater than a sum of throughputs for the plurality of service providers achievable when the virtual precoder matrices are applied to the signals.