A mathematical investigation published in Fixed Point Theory and Algorithms for Sciences and Engineering demonstrates that fractional stochastic evolution equations governed by the Hilfer derivative can achieve approximate controllability in the mean-square sense when driven by multivalued control operators. Researchers Azmat Ullah Khan Niazi, Ammar Alsinai, and Sadia Yasmeen formulated a theoretical framework addressing dynamical systems operating under the Hilfer fractional derivative of order between one and two and type varying from zero to one. By integrating the properties of multivalued maps with fixed-point methodologies, the authors established satisfactory conditions under which stochastic systems subjected to random perturbations can be steered arbitrarily close to any targeted terminal state in an infinite-dimensional space. The analysis directly resolves challenges stemming from memory effects and non-unique control trajectories by combining operator semigroup theory with specialized fractional calculus techniques.
Dynamical processes exhibiting hereditary properties and prolonged memory retention frequently challenge standard integer-order differential formulations, prompting the adoption of fractional-order differential equations. In practical physical and engineering settings, these dynamical structures are regularly subjected to unpredictable fluctuations, requiring stochastic differential equations that integrate randomness alongside memory parameters. While earlier investigations focused predominantly on deterministic models or stochastic formulations governed by Caputo-type derivatives, the Hilfer fractional derivative provides a broader mathematical architecture. Because the Hilfer derivative naturally interpolates between the classical Riemann-Liouville derivative when the type parameter equals zero and the Caputo formulation when the type parameter equals one, establishing controllability within this generalized structure resolves open questions across functional differential inclusions and hereditary stochastic systems.
To rigorously represent the state of the system over time, the team constructed a mild solution formulation grounded in operator theory on Banach spaces. The mathematical system incorporates an infinitesimal generator of a cosine operator family alongside bounded linear input mappings and nonlinear drift terms that satisfy defined Lipschitz and growth conditions. Stochastic excitation is modeled through an infinite-dimensional Wiener process defined on a complete filtered probability space with a nuclear covariance operator possessing finite trace. Mild solutions to the fractional stochastic differential inclusion were represented through integral equations featuring cosine operator families and Mainardi Wright kernel representations. These kernels translate fractional diffusion dynamics into concrete analytical operators, allowing researchers to evaluate state transitions across compact time intervals while preserving continuity and boundedness across the underlying function spaces.
The management of control inputs in this architecture requires addressing non-unique and multivalued input sets. The authors formalized control operators as multivalued maps belonging to the space of Carathéodory multifunctions with non-empty, bounded, closed, and convex values. Because multiple admissible controls can correspond to any given system state, the investigation applied measurable selection theorems to extract verifiable continuous trajectories from the multivalued feedback maps. This selection process ensures that the resulting operator graphs remain closed within continuous function spaces, establishing the functional foundation required to apply topological fixed-point principles to the combined stochastic differential inclusion.
The mathematical verification of approximate controllability relies on introducing a regularized controllability operator parameterized by a positive scalar beta. The authors constructed an explicit control function utilizing the adjoint of the fractional evolution operator combined with the beta-regularized resolvent operator. By substituting this control representation into the mild solution equation, the controllability problem was transformed into finding a fixed point for a multi-stage operator on continuous Banach spaces. The researchers systematically confirmed that the associated operator preserves convexity, maps bounded closed subsets into themselves under explicit parameter bounds, and transforms bounded sets into equicontinuous families of continuous stochastic processes.
Demonstrating equicontinuity and compactness required breaking down the trajectory differences across incremented time steps into seventeen distinct analytical components. Using uniform bounds derived for the cosine and sine operator families, the properties of the Mainardi Wright function, and the compactness of the underlying operator semigroup, the team proved that mean-square trajectory variations diminish to zero as temporal increments vanish. Furthermore, by constructing approximating subsets based on uniform operator bounds and invoking the classical Arzelà -Ascoli theorem, the multi-stage operator was shown to possess relatively compact ranges and a closed graph, thereby fulfilling the rigorous criteria of fixed-point theorems for upper semi-continuous multivalued mappings.
With the existence of mild solutions confirmed, the researchers evaluated the asymptotic behavior of the terminal system state as the regularization parameter beta approaches zero from above. Utilizing the stochastic Fubini theorem alongside the Lebesgue dominated convergence theorem, the team analyzed the mean-square distance between the terminal system trajectory and an arbitrary target state in the probability space. Because the regularized operator term approaches zero strongly as beta vanishes, and because the fractional evolution operators maintain uniform compactness, the expected squared error between the actual terminal state and the desired target can be made arbitrarily small, formally verifying approximate controllability in the mean-square sense.
To validate the operational relevance of the theoretical framework, the authors provided an explicit two-dimensional numerical application governed by a Hilfer derivative of order three-halves and type one-half, yielding an associated fractional integral parameter of one-fourth. The example incorporated a two-by-two system matrix alongside bounded input and noise projection vectors, state-dependent hyperbolic tangent nonlinearities, and nonlocal initial conditions evaluated across discrete interior time points. Under established Kalman controllability criteria for the deterministic pair, combined with sufficiently small noise and Lipschitz constants, the analysis confirmed that admissible feedback selections maintain mean-square error bounds below any predetermined threshold, corroborating the formal proofs through simulation trajectories of the stochastic state and control inputs.
The findings provide a foundational theoretical framework for controlling complex fractional evolution inclusions subjected to external stochastic noise and memory constraints. By unifying Hilfer fractional derivatives, cosine operator families, multivalued measurable selection, and regularized resolvent operators, the research establishes explicit criteria ensuring that stochastic systems can be driven to desired targets without requiring exact algebraic inversion of infinite-dimensional operators. The authors note that the underlying mathematical data supporting these theoretical derivations and simulation examples remain available upon reasonable request from the corresponding author.
In their comprehensive review of existing mathematical literature, the investigators highlighted how previous analytical research established key building blocks while leaving critical gaps. Early foundational works investigated approximate controllability across fractional neutral stochastic systems with infinite delay utilizing Mainardi functions, while related efforts examined delay integro-differential systems and neutral differential evolution equations. Other studies introduced methods for fractional evolution inclusions with order parameters strictly between one and two, addressing both infinite delay dynamics and hemivariational inequalities in abstract spaces. Related research teams analyzed stochastic differential inclusions driven by fractional Brownian motion in finite dimensions, alongside semilinear functional differential inclusions and fractional neutral equations featuring nonlocal boundary constraints. Subsequent developments extended these analytical tools to nonlocal impulsive systems, fractional sectorial operators in Banach spaces, and systems with Poisson jumps. However, the authors observed that these earlier models predominantly addressed deterministic dynamics or relied on Caputo-type derivatives, motivating the present comprehensive treatment of Hilfer-type stochastic evolution inclusions featuring multivalued control operators.
To establish the analytical infrastructure necessary for resolving mild solutions, the researchers articulated several precise operator properties and fractional calculus definitions. Within this structure, the fractional integral of arbitrary order incorporates convolution formulations alongside the classical Euler gamma function, while the Dirac measure represents conditions where the integral order reaches zero. The left Hilfer fractional derivative couples standard differentiation with lower-order fractional integration through parameters governing both order and type. Under this formulation, setting the type parameter to zero recovers the classical Riemann-Liouville fractional derivative, whereas selecting a type value of one reproduces the Caputo derivative. The mathematical framework also incorporates cosine and sine operator families generated by an infinitesimal operator, establishing uniform bounds where operator norms scale exponentially relative to time. Furthermore, resolvent operators mapping across the underlying Banach algebra provide integral representations essential for defining fractional evolution operators and tracking trajectory bounds across compact intervals.
The mathematical proof establishing mild solutions also relies heavily on structured hypotheses regarding system non-linearities and multivalued mappings. Under these formal assumptions, the nonlinear function satisfies mean-square growth conditions bounded by a Lebesgue-integrable coefficient function, ensuring controlled Lipschitz-type variance across state variables. Simultaneously, the multivalued control map is required to produce nonempty, closed, bounded, and convex selector sets belonging to the square-integrable function space, with second-order norms bounded by an integrable auxiliary function. The analytical existence theorem demonstrates that when a composite inequality linking terminal time, operator bounds, and growth parameters remains strictly below unity, the multi-stage trajectory mapping possesses a non-empty convex domain and maps bounded sets into equicontinuous subsets. This structure enables the application of Krasnoselskii-type and fixed-point theorems for upper semi-continuous multivalued maps, confirming the existence of at least one mild solution trajectory.
The authors illustrated their analytical findings using an explicit simulation system where the state matrix features off-diagonal couplings and negative trace parameters, paired with a two-dimensional vector steering input. The nonlocal initial configuration evaluates state distributions across two interior time values, while the initial derivative remains zero. Admissible control selections were restricted to bounded intervals scaling proportionally with the instantaneous state norm, ensuring that the control values remain physically admissible across all time steps. By establishing Kalman rank controllability for the linear pair and selecting sufficiently small constants for the nonlinear perturbations, the authors verified that the terminal state reaches within an arbitrary error threshold of any target point in two-dimensional Euclidean space. These analytical proofs confirm that Hilfer-governed stochastic differential inclusions achieve mean-square approximate controllability under explicit mathematical conditions.
Subject of Research: Approximate controllability of fractional stochastic differential equations with Hilfer derivatives
Article Title: Approximate controllability of fractional stochastic differential equations with Hilfer derivative: insights from multivalued maps and fixed-point method
Article References: Niazi, A. U. K., Alsinai, A., & Yasmeen, S. (2026). Approximate controllability of fractional stochastic differential equations with Hilfer derivative: insights from multivalued maps and fixed-point method. Fixed Point Theory and Algorithms for Sciences and Engineering, 2026(1), Article 14. https://doi.org/10.1186/s13663-026-00825-9
Image Credits: AI Generated
DOI: 10.1186/s13663-026-00825-9
Keywords: approximate controllability, fractional stochastic differential equations, Hilfer derivative, multivalued maps, fixed point method, stochastic evolution equations, cosine operator families, Wright functions, Mainardi kernel, Banach space, mild solutions, Wiener process, Krasnoselskii fixed point theorem, Arzela-Ascoli theorem, regularized controllability operator, stochastic analysis, Sobolev-type systems, Kalman controllability
News Source: Denise Maddox. (November 19, 2026). Hilfer Fractional Stochastic Systems Achieve Approximate Controllability via Multivalued Control Maps. Scienmag.



