Dynamics and Voting-Rule Interaction Concept
The thesis rests on a closed feedback loop: rule choice changes election outcomes, outcomes change incentives and learning, and those changes reshape later elections.
1. Closed-Loop Dynamic
The implemented loop links rules to participation through multiple channels:
- Rule aggregates mixed ballots into elected ordering.
- Elected ordering determines quality-gate sign and reward structure.
- Reward and fee components update participation learning signals.
- Participation composition in later steps changes ballot pools.
- Election outcomes shape later grid state (lagged mutation), which affects dissatisfaction and altruistic-vote probability.
This creates endogenous temporal dynamics where rule differences can accumulate.
2. Power-Struggle vs Quality-Alignment Tension
The model intentionally allows two directional pressures:
- self-regarding ballots that push toward group preference power
- altruistic ballots that use distributed knowledge to track quality pressure
Rule aggregation behavior determines how these pressures are combined each step.
3. Satisfying-Region Attractor Intuition (Design-Level)
If the system reaches a region where many agents are relatively satisfied, altruistic-vote probability tends to increase, which can improve quality-gate alignment and reduce destabilizing punishment regimes.
At the same time, self-regarding pressure can prevent convergence or create lock-in/polarized movement.
In the thesis, this attractor logic is treated as design intuition and exploratory expectation rather than as a primary confirmatory hypothesis.