Dynamic Graph Rewiring in Decentralized Exchange Networks

Dynamic Graph Rewiring in Decentralized Exchange Networks: A Study of MEV-Induced Liquidity Fragmentation. Nitish Kumar. Presented at the Nanyang Blockchain Conference (NBC ’26), NTU Singapore, 2026.

What the paper studies

The paper looks at how Maximal Extractable Value (MEV) extraction changes the shape of decentralised exchange (DEX) liquidity networks. Liquidity across DEX pools is modelled as a graph, and the question is how MEV activity rewires that graph over time, including whether liquidity fragments into separate clusters.

Method

The public repository describes "a hybrid approach combining on-chain aggregate statistics with calibrated simulation":

• Data: 500 calibrated data windows.
• Calibration sources: ZeroMEV API, Google BigQuery crypto_ethereum, Dune Analytics and Flashbots MEV-Boost relay data.
• Network metrics: modularity, spectral gap and effective connectivity.
• Figures: modularity evolution over 5M blocks; a phase transition fitted with a logistic curve; spectral gap evolution.
• Tooling: Python, NumPy, SciPy, pandas and matplotlib; the paper is typeset in LaTeX (ACM sigconf).

The repository does not publish numeric results, so this page reports only the method and outputs it states.

Code

The simulation pipeline and analysis code are public under the MIT licence at github.com/technitish9123/mev-liquidity-fragmentation.

@inproceedings{kumar2026dynamic,
  title     = {Dynamic Graph Rewiring in Decentralized Exchange Networks: A Study of MEV-Induced Liquidity Fragmentation},
  author    = {Kumar, Nitish},
  booktitle = {Proceedings of the Nanyang Blockchain Conference (NBC '26)},
  year      = {2026}
}