🧠 Episodic Memory • Ebbinghaus Decay • Mem0 & Zep • Recency vs Importance
AI Agent Memory Decay & Retention Simulator
Model how autonomous agents prioritize past interactions. Simulate exponential decay curves over elapsed time, boosted by reinforcement recall and semantic importance weights.
Memory Node Properties
Elapsed Time (Hours) 48 hours (2 days)
Semantic Importance Weight (1-10) 8 / 10
Reinforcement Recall Count 3 recalls
💡 Generative Agent Formula
Score = α × Recency(e-λt) + β × Importance + γ × Relevance. Frequently accessed memories reset their decay half-life.
Memory Retention Status
Active in Working ContextRetention Score
0.86
Decay Half-Life
46 hrs
Context Action
Inject
Memory Strength Over Time 86%
Agent Retrieval Vector JSON Payload