Constrained graph generation: Preserving diameter and clustering coefficient simultaneously

Researchers Dávid Ferenczi and Alexander Grigoriev have published a breakthrough paper on arXiv titled 'Constrained graph generation: Preserving diameter and clustering coefficient simultaneously,' addressing a fundamental challenge in network science. Generating graphs that simultaneously satisfy multiple structural constraints like diameter (the longest shortest path) and clustering coefficient (a measure of local connectivity) has proven exceptionally difficult, with traditional methods failing to explore the full space of feasible networks.

The team's novel solution combines Ant Colony Optimization (ACO) with Markov-chain Monte Carlo (MCMC) sampling in a two-step hybrid framework. First, a layered ACO heuristic performs guided global search to locate valid graphs with prescribed properties. Second, these ACO-designed graphs serve as structurally distinct seed states for an MCMC rewiring algorithm. Experiments across varying edge densities and constraint regimes revealed stark differences: standard MCMC samplers remained trapped in isolated subsets, while the hybrid approach successfully bridged disconnected configuration landscapes.

Using spectral distance of the normalized Laplacian to quantify structural diversity, the researchers demonstrated their method generates a 'vastly richer and structurally diverse set of valid graphs.' This breakthrough has significant implications for synthetic network generation, where realistic test networks must preserve multiple real-world properties simultaneously. The work addresses severe ergodicity breaking that has plagued traditional samplers, offering a practical solution for researchers and engineers needing constrained graph models for testing algorithms, simulating social networks, or creating benchmark datasets.