EvoSpark: Endogenous Interactive Agent Societies for Unified Long-Horizon Narrative Evolution

2026-04-14

AI Summary

This work addresses the challenges of redundant social memory and disconnection between narrative and spatial logic in long-term storytelling by large language model–driven multi-agent systems, stemming from generative stochasticity. To resolve these issues, we propose an endogenous interactive agent society framework that integrates hierarchical narrative memory, generative mise-en-scène, and an emergent role anchoring protocol. A key innovation is the introduction of a unified narrative operation engine that transforms stochastic generation into persistent character identities and a coherent story world. Central to this approach is the character socio-evolutionary substrate—a dynamic cognitive mechanism that continuously reconciles historical inconsistencies to ensure consistent co-evolution among characters, locations, and plotlines. Experimental results demonstrate that our method significantly outperforms baseline approaches across multiple paradigms, enabling logically consistent and expressively rich narrative generation over extended time spans.

Technology Category

• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Humans and AI: Game Design — Procedural Content Generation & Storytelling
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Realizing endogenous narrative evolution in LLM-based multi-agent systems is hindered by the inherent stochasticity of generative emergence. In particular, long-horizon simulations suffer from social memory stacking, where conflicting relational states accumulate without resolution, and narrative-spatial dissonance, where spatial logic detaches from the evolving plot. To bridge this gap, we propose EvoSpark, a framework specifically designed to sustain logically coherent long-horizon narratives within Endogenous Interactive Agent Societies. To ensure consistency, the Stratified Narrative Memory employs a Role Socio-Evolutionary Base as living cognition, dynamically metabolizing experiences to resolve historical conflicts. Complementarily, Generative Mise-en-Scène mechanism enforces Role-Location-Plot alignment, synchronizing character presence with the narrative flow. Underpinning these is the Unified Narrative Operation Engine, which integrates an Emergent Character Grounding Protocol to transform stochastic sparking into persistent characters. This engine establishes a substrate that expands a minimal premise into an open-ended, evolving story world. Experiments demonstrate that EvoSpark significantly outperforms baselines across diverse paradigms, enabling the sustained generation of expressive and coherent narrative experiences.

Problem

• endogenous narrative evolution
• social memory stacking
• narrative-spatial dissonance
• long-horizon simulation
• multi-agent systems

Innovation

• Endogenous Interactive Agent Societies
• Stratified Narrative Memory
• Generative Mise-en-Scène
• Unified Narrative Operation Engine
• Emergent Character Grounding