Weighted Quorum Decision-Making in Distributed Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing decision-making algorithms in distributed networks face challenges such as dependence on network stability, potential obsolescence of decisions, and high load on central nodes, particularly in centralized and distributed algorithms, and lack of versatility in hybrid algorithms.
Innovation Solution
A hybrid, weighted decision-making process that allows for centralized, distributed, or hybrid decision-making based on voter influence, with a quorum principle, where voters can provide confirmations, disagreements, or neutral opinions, and a timer ensures timely decisions, independent of transport network quality, and allows for dynamic selection of voters based on proximity and capacity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a centralized algorithm is used where supervisor A makes decisions for all nodes B, then supervisor A is fully aware of the network environment and nodes B can be simple machines, but the decision depends on network stability and may become outdated, and node A may become overloaded when multiple nodes request decisions simultaneously
Solution Approach 1:
The patent segments the centralized decision-making process into distributed decision-making units at each node. Instead of one supervisor A handling all decisions, each node B is equipped with local decision-making intelligence that can autonomously make decisions based on local environmental parameters, eliminating the bottleneck and latency associated with centralized processing.
Solution Approach 2:
The patent transitions from a single-dimensional centralized hierarchy to a multi-dimensional distributed architecture where decision-making capability exists at multiple levels (individual nodes, groups of nodes, and supervisor nodes). This dimensional expansion allows simultaneous local and global decision-making, reducing latency while maintaining comprehensive awareness.
2Loss of energy
If a distributed algorithm is used where all decision-making information passes through transport network TN2, then few signals are exchanged and network load is low, but the decision depends on the stability of TN2 network and nodes B must be capable of calculating decisions
Solution Approach 1:
The patent applies local quality by enabling each node to make decisions based on local environmental parameters and local voting, rather than requiring all decisions to pass through the transport network. This reduces network load while maintaining reliability through local autonomy. Nodes only exchange decisions when needed, not continuous environmental data.
Solution Approach 2:
The patent introduces supervisor nodes as intermediaries between fully centralized and fully distributed models. Supervisor nodes coordinate decisions for groups of nodes, providing reliability oversight while allowing local autonomy. This intermediary layer ensures decision reliability without requiring constant network communication for all decisions.
3Adaptability or versatility
If a hybrid algorithm is used where some decisions are made centrally by supervisor A and others by nodes B, then fewer signals are needed and decisions based on distributed algorithms are not obsolete, but nodes A and B must be capable of calculating decisions and the decision may depend on either transport network
Solution Approach 1:
The patent implements dynamic decision-making where the algorithm type (centralized, distributed, or hybrid) is not fixed but adapts based on the specific decision context, environmental parameters, and network conditions. Each node can dynamically select the most appropriate decision-making approach for each situation, enhancing versatility while managing complexity through context-aware selection.
Solution Approach 2:
The patent creates a universal decision-making framework that can operate in multiple modes (centralized, distributed, hybrid) within the same system. All nodes are equipped with the capability to perform all three types of decisions, making the system multi-functional and adaptable to different scenarios without requiring separate systems for each decision type.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
The invention concerns a quorum decision-making process. It implements the propagation of a proposition in a distributed network of processing units with a view to a decision being taken by a so-called decider processing unit, each processing unit being provided with decision intelligence. This involves: a) the decider determining a proposition in response to a request, b) selecting a set of processing units as voters, c) defining a percentage weighting for each voter, d) defining a threshold when the response is a confirmation of a proposition, e) defining a second threshold when the response is a denial, f) triggering a timer and defining at least one processing time, g) transmitting said proposition with the processing time via the distributed network to all the voters, h) following this processing time, the decider gathers all the available responses and determines if one of the thresholds has been reached.