Dual-Model framework for CHIKV transmission modeling: ODE and Petri Net Analysis of the 2025 Foshan outbreak

This study addresses the 2025 Chikungunya outbreak in Foshan City, Guangdong Province, China, by constructing a dual-model framework based on Ordinary Differential Equations (ODE) and Petri Nets (PN) for comparative analysis of Chikungunya transmission dynamics and reproduction number estimation methods. The research employs SEICR (Susceptible-Exposed-Infectious-Chronic-Recovered) compartmental modeling to compare two formal representations under matched epidemiological assumptions, and evaluates the timing of epidemic control measures through a three-phase intervention fitting protocol. Model validation results show that both models achieve root mean square errors (RMSE) of 30.98 (ODE) and 31.05 (PN), mean absolute errors (MAE) of 15.57 and 15.78, and [Formula: see text] and 0.9498, respectively. Both models predict epidemic peaks at day 33 (406 cases), occurring 3 days earlier than the observed peak (432 cases), with a peak value error of 6.0%. Residual analysis reveals that negative residuals account for 71.4% (ODE) and 73.8% (PN) of the observation-window residuals, suggesting a structured overprediction pattern in descriptive diagnostics. Reproduction number analysis reveals that the initial transmission indicators are approximately 14.67 (ODE)/13.90 (PN), with effective values progressively decreasing through three intervention phases: 7.85/7.86 after Phase 1, 7.59/7.56 after Phase 2, and 0.059 in Phase 3, below the transmission threshold. An additional no-demography robustness check shows that removing demographic turnover changes total predicted cases by only 0.03%, suggesting that the remaining uncertainty lies mainly in omitted vector-side dynamics rather than in human-side demography. Sensitivity analysis indicates that the recovery rate (γ) is the most sensitive parameter affecting [Formula: see text] within this formulation, with a Sobol index of 0.9672, explaining 96.72% of total [Formula: see text] variation. This study provides a controlled comparison between ODE and Petri Net representations of the same epidemiological structure, offering a transparent comparative framework for outbreak fitting, intervention phase identification, and future extension toward explicit host-vector models.