Skip to content

Oracle Selection Algorithms

⚠️ STUB DOCUMENTATION: This file contains placeholder content and needs to be expanded with detailed oracle selection algorithm documentation.

How oracles are selected for dispute resolution requests.

Selection Criteria

Primary Factors

  • Reputation Score: Past performance and accuracy
  • Availability: Current online status and capacity
  • Oracle Class: Matching requested specialization
  • Fee Structure: Competitive pricing within limits

Weighted Selection Algorithm

Selection Weight = (Reputation^α) × Availability × Class_Match × Fee_Factor

Where: - α (alpha): Reputation weighting parameter (0-1000) - Higher alpha = more weight on reputation - Lower alpha = more randomized selection

Selection Types

Single Oracle (ReputationSingleton)

  • Selects one oracle using weighted random selection
  • Fastest resolution, lowest cost
  • Suitable for simple disputes

Multi-Oracle (ReputationAggregator)

  • Selects multiple oracles for consensus
  • Higher security through redundancy
  • Suitable for high-value disputes

Availability Management

Capacity Tracking

Oracles report their current capacity and maximum concurrent evaluations.

Load Balancing

System distributes requests to prevent oracle overload.

Fallback Selection

If preferred oracles are unavailable, system selects alternatives.

Implementation Details

Alpha Parameter Effects

  • α = 1000: Always select highest reputation oracle
  • α = 500: Balanced selection (recommended)
  • α = 0: Purely random selection (lowest fees)

For implementation examples: - ReputationKeeper Contract - Integration Examples - Reputation System Details