Paper: arXiv 2604.00556

Authors: Hongyang Yang, Yanxin Zhang, Yang She, Yue Xiao, Hao Wu, Yiyang Zhang, Jiapeng Hou, Rongshan Zhang

Abstract

Housing selection is a high-stakes and largely irreversible decision problem. We study housing consultation as a decision-support interface for housing selection. Existing housing platforms and many LLM-based assistants often reduce this process to ranking or recommendation, resulting in opaque reasoning, brittle multi-constraint handling, and limited guarantees on factuality. We present HabitatAgent, the first LLM-powered multi-agent architecture for end-to-end housing consultation. HabitatAgent comprises four specialized agent roles: Memory, Retrieval, Generation, and Validation. The Memory Agent maintains multi-layer user memory through internal stages for constraint extraction, memory fusion, and verification-gated updates; the Retrieval Agent performs hybrid vector–graph retrieval (GraphRAG); the Generation Agent produces evidence-referenced recommendations and explanations; and the Validation Agent applies multi-tier verification and targeted remediation. Together, these agents provide an auditable and reliable workflow for end-to-end housing consultation. We evaluate HabitatAgent on 100 real user consultation scenarios (300 multi-turn question–answer pairs) under an end-to-end correctness protocol. A strong single-stage baseline (Dense+Rerank) achieves 75% accuracy, while HabitatAgent reaches 95%.

Complexity vs Empirical Score

  • Math Complexity: 2.0/10
  • Empirical Rigor: 6.0/10 (capped from 6.5: no out-of-sample test; see penalty flags)
  • Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper presents an applied LLM multi-agent framework (HabitatAgent) for housing selection with strong pragmatic design and a clear evaluation setup. While mathematical complexity is minimal, the empirical design evaluates real multi-turn scenarios effectively against standard baselines.

Research Flowchart

  flowchart TD
    A["Research Goal: End-to-End, Fact-Based Housing Consultation"] --> B["Data Input: 100 Real User Scenarios <br/>(300 Multi-Turn QA Pairs)"]
    
    subgraph C["HabitatAgent Architecture"]
        C1["1. Memory Agent: Constraint Extraction & Fusion"] --> C2["2. Retrieval Agent: Hybrid Vector-Graph Retrieval (GraphRAG)"]
        C2 --> C3["3. Generation Agent: Evidence-Referenced Recommendations"]
        C3 --> C4["4. Validation Agent: Multi-Tier Verification & Remediation"]
    end

    B --> C
    C --> D["Evaluation: End-to-End Correctness Protocol"]
    
    D --> E["Key Outcome: 95% Accuracy vs. <br/>75% Single-Stage Baseline (Dense+Rerank)"]

Cited References (144)

  1. Options, futures and other derivatives/John C. Hull. (2009)
  2. Fuzzy adaptive decision-making for boundedly rational traders in speculative stock markets (2010)
  3. Online portfolio selection strategy based on combining experts’ advice (2017)
  4. An intelligent hybrid trading system for discovering trading rules for the futures market using rough sets and genetic algorithms (2017)
  5. Learning to trade via direct reinforcement (2001)
  6. Deep learning for finance: deep portfolios (2017)
  7. Portfolio theory and capital markets (1970)
  8. Practical deep reinforcement learning approach for stock trading (2018)
and 136 more
  1. Optimistic bull or pessimistic bear: Adaptive deep reinforcement learning for stock portfolio allocation (2019)
  2. Multi-agent deep reinforcement learning for liquidation strategy analysis (2019)
  3. Introduction to discrete-event simulation and the simpy language (2008)
  4. TensorLayer : A versatile library for efficient deep learning development (2017)
  5. Benchmarking deep reinforcement learning for continuous control (2016)
  6. Continuous control with deep reinforcement learning (2016)
  7. Mastering the game of Go with deep neural networks and tree search (2016)
  8. Reinforcement learning for control: Performance, stability, and deep approximators (2018)
  9. Optimal execution of portfolio transactions (2001)
  10. Learning from delayed rewards (1989)
  11. Multi-agent actor-critic for mixed cooperative-competitive environments (2017)
  12. Dynamic programming and markov processes. (1960)
  13. PyBrain (2010)
  14. Human-level control through deep reinforcement learning (2015)
  15. CSI 300 (2017) — link
  16. Deep hedging: Hedging derivatives under generic market frictions using reinforcement learning (2019)
  17. Garage: A toolkit for reproducible reinforcement learning research (2019)
  18. Pyfolio: A toolkit for Portfolio and risk analytics in Python (2019)
  19. “Personae” (2018) — link
  20. Investment science (1997)
  21. Experience replay optimization (2019)
  22. Hang Seng Index and Sub-indexes (2020) — link
  23. SSE 180 Index Methodology (2018) — [link](http://www.sse.com.cn/market/sseindex/indexlist/indexdetails/indexmethods/c/Index Handbook_EN_SSE 180.pdf)
  24. Shenzhen Component (2019) — link
  25. A Markovian decision process (1957)
  26. Proximal policy optimization algorithms (2017) — link
  27. Soft actor-critic: Off-policy maximum entropy deep reinforcement learning with a stochastic actor (2018)
  28. Asynchronous methods for deep reinforcement learning (2016)
  29. Addressing function approximation error in actor-critic methods (2018)
  30. Mean-variance investing (August 10, 2012)
  31. OpenAI Gym (2016) — link
  32. Reinforcement learning in financial markets - a survey (2018)
  33. Reinforcement learning: An introduction (2018)
  34. Deep Reinforcement Learning with Double Q-Learning (2016)
  35. Dueling Network Architectures for Deep Reinforcement Learning (2016)
  36. Policy gradient methods for reinforcement learning with function approximation (2000)
  37. Asynchronous methods for deep reinforcement learning (2016)
  38. Trust region policy optimization (2015)
  39. Twin-Delayed DDPG : A Deep Reinforcement Learning Technique to Model a Continuous Movement of an Intelligent Robot Agent (2019)
  40. RLlib : Abstractions for Distributed Reinforcement Learning
  41. OpenAI Baselines (2017)
  42. Stable Baselines (2018)
  43. Dopamine: A research framework for deep reinforcement learning (2018)
  44. Passive Investment Strategies and Efficient Markets (2003)
  45. Exchange Traded Funds (ETF): History, Mechanism, Academic Literature Review and Research Perspectives (2016)
  46. Skulls, financial turbulence, and risk management (2010)
  47. LIFT : Reinforcement Learning in Computer Systems by Learning From Demonstrations (2018) — link
  48. Horizon: Facebook’s Open Source Applied Reinforcement Learning Platform (2018) — link
  49. Deep Reinforcement Learning for Automated Stock Trading: An Ensemble Strategy (2020)
  50. Reinforcement Learning in Stock Trading (2019)
  51. Portfolio management using reinforcement learning (2016)
  52. Financial trading as a game: A deep reinforcement learning approach (2018) — link
  53. Sentiment and Knowledge Based Algorithmic Trading with Deep Reinforcement Learning (2020)
  54. Performance functions and reinforcement learning for trading systems and portfolios (1998)
  55. Deep reinforcement learning for trading (2020)
  56. Deep Direct Reinforcement Learning for Financial Signal Representation and Trading (2017)
  57. Reinforcement Learning For Automated Trading (2016)
  58. Algorithmic Trading using Sentiment Analysis and Reinforcement Learning (2016)
  59. Q-Learning-Based Financial Trading Systems with Applications (2014)
  60. Deep Reinforcement Learning in High Frequency Trading (2018)
  61. A Deep Reinforcement Learning Approach for Automated Cryptocurrency Trading (2019)
  62. Deep Reinforcement Learning in Cryptocurrency Market Making (2019)
  63. Cryptocurrency portfolio management with deep reinforcement learning (2017)
  64. Bitcoin Is Volatile! Isn’t that Right? (2014)
  65. A Deep Reinforcement Learning Framework for the Financial Portfolio Management Problem (2017)
  66. Deep Hedging of Derivatives Using Reinforcement Learning (2019)
  67. Application of deep reinforcement learning on automated stock trading (2019)
  68. Learning to Trade with Deep Actor Critic Methods (2018)
  69. Profitable momentum trading strategies for individual investors (2015)
  70. FinRL: A Deep Reinforcement Learning Library for Automated Stock Trading in Quantitative Finance (2020) — link
  71. TensorTrade: An open source Python framework for trading algorithms using reinforcement learning (2019)
  72. Backtrader: A feature-rich Python framework for backtesting and trading (2015)
  73. Zipline: A Pythonic Algorithmic Trading Library (2014)
  74. bt: Flexible backtesting for Python (2014)
  75. vectorbt: Portfolio optimization and backtesting on pandas/NumPy (2020)
  76. QSTrader: A Python-based open-source modular backtesting framework (2014)
  77. Alpaca API Documentation: Paper Trading (2025)
  78. Time Series Momentum (2012)
  79. Trading Systems and Methods (1998)
  80. Financial Modeling Prep (2026)
  81. Time series momentum (2012)
  82. Trading Systems and Methods (1998)
  83. The Probability of Backtest Overfitting (2017) — link
  84. The Investor Fear Gauge (2000) — link
  85. FinGPT: Open-Source Financial Large Language Models — → 2306.06031
  86. Qlib: An AI-Oriented Quantitative Investment Platform (2020)
  87. TradeMaster: A Holistic Quantitative Trading Platform Empowered by Reinforcement Learning (2023)
  88. FinRL-Meta: Market Environments and Benchmarks for Data-Driven Financial Reinforcement Learning (2022)
  89. Lean: Algorithmic Trading Engine (2024)
  90. Ai realtor: Towards grounded persuasive language generation for automated copywriting (2025) — link
  91. Real estate with AI: An agent based on LangChain (2024)
  92. A survey on LLM -based multi-agent systems: workflow, infrastructure, and challenges (2024) — link
  93. CAMEL: Communicative Agents for ``Mind’’ Exploration of Large Language Model Society (2023) — link
  94. MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework (2023) — link
  95. AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation (2023) — link
  96. MedAgents: Large Language Models as Collaborators for Zero-shot Medical Reasoning (2024)
  97. MDAgents: An Adaptive Collaboration of LLM s for Medical Decision Making (2024) — link
  98. LegalAgentBench: Evaluating LLM Agents in Legal Domain (2025)
  99. FinMem: A Performance-Enhanced LLM Trading Agent With Layered Memory and Character Design — → 2311.13743
  100. TradingAgents: Multi-Agents LLM Financial Trading Framework — → 2412.20138
  101. FinRobot: AI Agent for Equity Research and Valuation with Large Language Models — → 2411.08804
  102. REAL: Benchmarking Abilities of Large Language Models for Housing Transactions and Services (2025) — link
  103. FairHome: A Fair Housing and Fair Lending Dataset (2024) — link
  104. The Analytic Hierarchy Process (1980)
  105. Using the Analytic Hierarchy Process in House Selection (1994)
  106. Using Goal Programming and the Analytic Hierarchy Process in House Selection (1995)
  107. Real estate with AI : An agent based on LangChain , GPT and RAG for professional real estate market value estimation (2024)
  108. Recommender systems in real estate: a systematic review (2025)
  109. A Survey on LLM -Based Multi-Agent Systems: Workflow, Infrastructure, and Challenges (2024) — link
  110. Multi-Agent Collaboration Mechanisms: A Survey of LLM s (2025) — link
  111. MACRec : A Multi-Agent Collaboration Framework for Recommendation (2024) — link
  112. Large Language Model Based Multi-agents: A Survey of Progress and Challenges (2024) — link
  113. Retrieval-Augmented Generation with Graphs ( GraphRAG ) (2025) — link
  114. GraphRAG
  115. Recommender Systems in the Real Estate Market—A Survey (2021) — link
  116. Recommender Systems in Real Estate: A Systematic Review (2025) — link
  117. Conversational Recommendation: A Grand AI Challenge (2022) — link
  118. LLM -based Conversational Recommendation Agents with Collaborative Verbalized Experience (2025) — link
  119. Detecting Hallucinations in Large Language Models Using Semantic Entropy (2024) — link
  120. LangGraph : Agent Orchestration Framework for Reliable AI Agents
  121. MemRec : Collaborative Memory-Augmented Agentic Recommender System (2026) — link
  122. LLM agents for smart city management: Enhancing decision support through multi-agent AI systems (2025)
  123. Real-GPT: efficiently tailoring LLMs for informed decision-making in the real estate industry (2025)
  124. EXPRESS: An LLM-generated explainable property valuation system with neighbor imputation (2025) — link
  125. On the performance of LLMs for real estate appraisal (2025)
  126. The Agentic Frontier: Artificial Intelligence in Real Estate (2025)
  127. SimCity: Multi-Agent Urban Development Simulation with Rich Interactions (2025) — link
  128. A behavioural agent-based model for housing markets: impact of financial shocks in the UK (2024)
  129. RE-RecSys: An End-to-End system for recommending properties in Real-Estate domain (2024) — link
  130. FinRobot: An Open-Source AI Agent Platform for Financial Applications using Large Language Models — → 2405.14767
  131. Bloomberggpt: A large language model for finance (2023) — link
  132. Instruct-FinGPT: Financial Sentiment Analysis by Instruction Tuning of General-Purpose Large Language Models (2023)
  133. Enhancing Financial Sentiment Analysis via Retrieval Augmented Large Language Models — → 2310.04027
  134. FinGPT: enhancing sentiment-based stock movement prediction with dissemination-aware and context-enriched LLMs (2025)
  135. FinGPT: Instruction Tuning Benchmark for Open-Source Large Language Models in Financial Datasets (2023)
  136. Enhancing Investment Analysis: Optimizing AI-Agent Collaboration in Financial Research (2024)

6 of 144 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.