Distributed Task Scheduling via Game Theory Consensus

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing distributed task scheduling systems face challenges with computationally costly auction-based bid generation techniques that are not suitable for real-time scheduling, and heuristic approaches often fail to find optimal solutions, with the existence of pure strategy Nash equilibrium not being guaranteed.

Innovation Solution

A processor-implemented method for game theory-based real-time distributed dynamic task scheduling among agents, involving self-allocation of tasks based on predefined constraints, multi-agent Markov decision processes, and pure strategy correlated equilibrium to achieve consensus among agents, ensuring tasks are completed within deadlines with minimized penalties.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If auction-based bid generation technique is used in distributed task scheduling, then reliability is improved, but computational complexity increases and real-time scheduling becomes unsuitable

Engineering Contradiction:
ImprovereliabilityVSAvoidcomputational complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the task scheduling problem by dividing tasks into different priority levels (high priority and low priority tasks) and handling them through different mechanisms. High priority tasks use centralized scheduling for reliability, while low priority tasks use distributed self-allocation to reduce computational complexity. This segmentation resolves the contradiction by applying different strategies to different parts of the system.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces dynamic priority adjustment where task priorities can change based on system state and agent capabilities. The scheduling algorithm dynamically adapts between centralized and distributed modes based on real-time conditions, allowing the system to maintain reliability when needed while reducing computational complexity during normal operation.

Inventive Principle:
Principle #15Dynamics

2Device complexity

If heuristic approaches are used for task scheduling, then computational complexity is reduced, but optimality of solutions deteriorates

Engineering Contradiction:
Improvecomputational complexityVSAvoidoptimality
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent changes the parameter of task priority from static to dynamic, allowing the system to adjust priorities based on agent capabilities and system state. This enables the heuristic scheduling to achieve better optimality by adapting to changing conditions rather than following fixed rules, while still maintaining lower computational complexity than exact methods.

Inventive Principle:
Principle #35Parameter changes

3Device complexity

If centralized task scheduling is used, then computational complexity is reduced, but system reliability deteriorates due to single point failure

Engineering Contradiction:
Improvecomputational complexityVSAvoidreliability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The patent segments the scheduling function between centralized and distributed components. The centralized controller handles high priority tasks and overall coordination, while agents independently manage low priority tasks. This segmentation eliminates the single point of failure issue while keeping computational complexity manageable through clear division of responsibilities.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces task priority as an intermediary mechanism that mediates between centralized control and distributed autonomy. High priority tasks require centralized coordination through the intermediary priority system, while low priority tasks can be autonomously allocated, thus maintaining reliability without requiring complete centralization.

Inventive Principle:
Principle #24Intermediary (Mediator)

4Reliability

If distributed task scheduling is used, then system reliability is improved, but computational complexity increases making real-time scheduling unsuitable

Engineering Contradiction:
ImprovereliabilityVSAvoidreal-time performance
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent segments tasks into high priority and low priority categories, applying different scheduling strategies to each. High priority tasks use centralized real-time scheduling for critical operations, while low priority tasks use distributed self-allocation for non-critical operations. This segmentation enables real-time performance for essential functions while maintaining distributed reliability for overall system operation.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4002234B1System and method for a real-time distributed dynamic task scheduling
Publication Date: 2025.06.25 TATA CONSULTANCY SERVICES LTD
  • EP4002234B1 patent drawingFigure 1
  • EP4002234B1 patent drawingFigure 2
  • EP4002234B1 patent drawingFigure 3

AI summary

An auction-based bid generation technique is a NP-hard problem and is not suitable for real-time scheduling of multi-agents. The embodiments thus provide a system and method for scheduling a set of tasks among a plurality of agents. Herein, agents self-allocate tasks among themselves dynamically in a distributed fashion, following an ordered sequence of agent indexes. The motivation of following the ordered sequence of agent indexes is allowing each agent to select its best strategy once by exploiting the greedy characteristic of the agent. The preferred agent (based on the ordered sequence) self-allocates tasks among multiple based on the minimum L2 Norm between task attributes and agent attributes, results in a strategy. The strategy offered by a sequence needs to satisfy constraints. A heuristic reward function for each strategy is proposed. Based on these rewards, agents reach consensus by playing an exact potential game for scheduling tasks among the agents.