Mathematicians have built a fractional-order epidemic model that tracks Nipah virus transmission simultaneously across human, pig, and bird populations, and paired it with an optimal control framework to identify the most effective intervention strategies. The study, published as open access research in Discover Analytics on 13 July 2026 by Sunday Oluwafemi Gbodogbe and Adedapo Chris Loyinmi, formulates the model using the Caputo fractional derivative, an operator that incorporates memory effects arising from past infection states, which classical integer-order models do not account for.
The motivation rests on the epidemiology of Nipah virus, a zoonotic pathogen the World Health Organization has classified as a priority pathogen requiring urgent research attention. Bats of the genus Pteropus serve as the primary natural reservoir, pigs act as intermediate hosts, and humans acquire infection through contaminated fruit, exposure to bat saliva or urine, contact with infected pigs, or human-to-human transmission in healthcare and family settings. No approved vaccine or targeted antiviral treatment exists, and patient management relies primarily on supportive care, with antiviral agents such as ribavirin showing limited effectiveness.
The model divides human and pig populations into susceptible, infected, treated, and recovered compartments, while the bird population is partitioned into susceptible and infected classes. Transmission occurs through standard-incidence forces of infection that explicitly represent cross-species pathways: birds transmit to pigs and humans, pigs transmit to humans and within their own population, and infected humans contribute to further human spread. Birds act as the primary reservoir in the model’s biological assumptions, pigs serve as intermediate hosts capable of passing infection to humans, and treatment reduces disease burden without conferring immediate immunity.
The Caputo fractional derivative of order between zero and one introduces a memory kernel that allows past states to influence current dynamics, capturing effects such as cumulative exposure, delayed disease progression, and behavioral adaptation. A uniform fractional order is assumed across all compartments for analytical consistency. The authors establish that solutions remain nonnegative and bounded within a biologically feasible region, prove existence and uniqueness of solutions via a fixed-point argument, and derive the disease-free equilibrium where all infected and treated compartments vanish.
Using the next-generation matrix method, the team computed the basic reproduction number as the maximum of three host-specific thresholds, one each for humans, pigs, and birds. Because the underlying matrix is upper triangular, its eigenvalues fall on the diagonal, yielding a closed-form expression in which transmission coefficients and contact modification parameters appear in the numerator and treatment, natural mortality, and disease-induced mortality rates in the denominator. Sensitivity analysis shows that transmission rates and contact modification parameters carry positive sensitivity indices of one, making them the dominant drivers of disease spread, while treatment and removal rates reduce transmission.
The stability analysis follows the fractional-order criterion that an equilibrium is locally asymptotically stable only when every eigenvalue of the linearized system satisfies an argument condition tied to the fractional order. The disease-free equilibrium is locally asymptotically stable when the reproduction number is below one and unstable when it exceeds one, with a Lyapunov-based argument extending this to global stability within the feasible region. When the reproduction number surpasses one, the authors show that a unique endemic equilibrium exists and can be locally stable under the same eigenvalue condition.
Numerical simulations using a predictor-corrector Adams-Bashforth-Moulton scheme with a time step of 0.1 reveal a striking role for memory. Across scenarios testing fractional orders of 0.8, 0.9, and the classical integer-order case of 1, decreasing the fractional order significantly delayed epidemic peaks and prolonged infection duration, with the strongest sensitivity appearing in the human and pig populations. The abstract reports that reducing the fractional order lowered peak infection levels by more than 40 percent in some scenarios, while the classical case exhibited faster epidemic growth and quicker decay. The bird population showed monotone decline across all orders, though with markedly slower decline for smaller fractional values.
The optimal control framework introduces four time-dependent interventions bounded between zero and one: preventive measures reducing transmission to humans, intensified treatment for infected humans, biosecurity measures in pigs, and intensified treatment for infected pigs. Applying Pontryagin’s Maximum Principle yields a Hamiltonian, an adjoint system, and explicit characterizations of the optimal controls in terms of costate variables and infection forces. The objective functional balances the epidemiological burden of infection in each host population against quadratic costs of implementing each control.
The simulation results carry a consistent practical message. Under optimal control, infected human and pig populations rapidly decline to near-zero levels in both disease-free and endemic settings, whereas uncontrolled dynamics show pronounced epidemic peaks followed by slow decay. Susceptible human and pig populations are better preserved under control, treated compartments attain substantially lower peaks, and phase portraits show controlled trajectories contracting toward the disease-free equilibrium while uncontrolled trajectories exhibit large persistent excursions. The optimal controls are strongest at early times and decrease thereafter, a front-loaded structure the authors interpret as evidence that early aggressive intervention is most effective.
The authors frame the work as filling a gap in which existing models either focused on single-host dynamics without memory effects or applied fractional-order models to single populations without cross-species spillover mechanisms. They caution that the model is general rather than fitted to specific outbreak data: parameter values are taken from existing literature on Nipah virus dynamics or chosen to be biologically reasonable where empirical estimates are unavailable, no empirical datasets were generated or analyzed, and future work is expected to incorporate spatial heterogeneity, stochastic effects, and data-driven parameter estimation. Within those limits, the study argues that memory effects and coordinated, early interventions targeting both human and animal populations jointly shape whether a zoonotic spillover event fades out or persists.
Subject of Research: Fractional-order mathematical modeling of zoonotic Nipah virus spillover and optimal control across human, pig, and bird populations
Article Title: Fractional modeling of zoonotic spillover with optimal control in a multi host system
Article References: Gbodogbe, S. O., & Loyinmi, A. C. (2026). Fractional modeling of zoonotic spillover with optimal control in a multi host system. Discover Analytics, 4(1), Article 28. https://doi.org/10.1007/s44257-026-00079-0
Image Credits: AI Generated
DOI: 10.1007/s44257-026-00079-0
Keywords: Nipah virus, zoonotic spillover, fractional calculus, Caputo derivative, optimal control, mathematical modeling, epidemiology, basic reproduction number, multi-host transmission, Pontryagin Maximum Principle, Fractional, modeling
News Source: Kristina Jarvis. (October 29, 2026). Memory Effects and Optimal Control Reshape Nipah Virus Spillover in Fractional Multi-Host Model. Scienmag.
