# Faina Kirillova Faina Mikhailovna Kirillova (born 1931) is a Soviet and Belarusian mathematician and one of the builders of the Minsk school of [[Optimal_control|optimal control]] — the branch of [[Control_theory|control theory]] that asks not merely whether a [[Dynamical_system|dynamical system]] can be steered, but how to steer it best against a cost functional. In a decades-long partnership with Rafail Gabasov at the Institute of Mathematics of the National Academy of Sciences of Belarus, of which she became a corresponding member, Kirillova helped move the field through its three postwar phases: the *qualitative* theory (existence of optimal controls, [[Observability|observability]], controllability), the anatomy of *singular* regimes where the classical necessary conditions go silent, and the *constructive* theory — algorithms meant to compute optimal [[Feedback|feedback]] in real time rather than admire it in closed form. The Gabasov–Kirillova monographs, translated and cited across the [[Cybernetics_in_the_Soviet_Union|Soviet]] and Western control communities alike, made Minsk one of the recognized capitals of [[Mathematical_optimization|optimization]]-based control and trained generations of Belarusian control scientists and [[Applied_mathematics|applied mathematicians]]. ## The problem optimal control solves An optimal-control problem couples an [[Ordinary_differential_equation|ordinary differential equation]] ẋ = f(x, u, t), a control u(·) confined to an admissible set, boundary conditions, and a performance index to be minimized — fuel, time, tracking error. The subject crystallized in the 1950s out of the classical [[Calculus_of_variations|calculus of variations]] under the pressure of problems the classical theory handled badly: bounded controls, bang-bang switching, trajectory constraints from [[Aerospace_engineering|aerospace]] guidance. Two great instruments arrived nearly simultaneously — Pontryagin's maximum principle in Moscow (published with his students in the late 1950s) and [[Richard_Bellman|Bellman]]'s [[Dynamic_programming|dynamic programming]] in the United States — with [[Kalman_filter|Kalman]]'s state-space [[Estimation_theory|filtering]] and controllability circle completing the modern kit around 1960. Soviet [[Engineering_cybernetics|engineering cybernetics]] gave this mathematics institutional priority, and the Minsk group formed inside that national effort as [[Cybernetics|cybernetics]] shed its pseudoscience label. ## Qualitative theory: what must be true before you compute Kirillova's early reputation rests on the qualitative side: conditions under which optimal controls exist at all, how attainable sets behave, and how [[Functional_analysis|functional-analytic]] machinery — weak convergence, separation of convex sets, duality — puts the maximum principle and existence theorems on one foundation. The Gabasov–Kirillova *Qualitative Theory of Optimal Processes* (Russian edition 1971; English translation 1976) systematized this program, treating controllability, observability, existence, and necessary conditions as one structure rather than a list of tricks, and extending the analysis to systems with time delay — functional [[Differential_equation|differential equations]] whose state is a function segment rather than a point and naive extensions of the [[Lyapunov_stability|stability]] and control theorems fail. The delay-system results mattered practically — transport lags are the rule in chemical plants and [[Process_engineering|process engineering]] — and theoretically, as an early systematic treatment of infinite-dimensional state in applied [[Control_engineering|control engineering]]. ## Singular controls: where the maximum principle goes quiet The maximum principle picks the control that maximizes the control Hamiltonian — the descendant of [[Hamiltonian_mechanics|Hamiltonian mechanics]] — along the optimal trajectory; a *singular* arc is a stretch where that maximization fails to determine the control — the switching function vanishes identically, and higher-order conditions (generalized Legendre–Clebsch inequalities) must adjudicate. Singular regimes are not pathology: minimum-fuel spacecraft arcs, optimal harvesting of biological [[Population_dynamics|populations]], and many process-control programs live exactly there. Gabasov and Kirillova's *Singular Optimal Controls* (1973) was among the first books devoted to the phenomenon, assembling order tests, junction conditions, and worked structure for problems where the textbook recipe returns silence — a contribution with the same flavor as the hard cases that [[Uncertainty|uncertainty]] and degeneracy pose elsewhere in [[Applied_mathematics|applied mathematics]]. ## The constructive turn: algorithms over formulas From the late 1970s the Minsk school argued a blunt thesis: a solution that cannot be computed under realistic constraints is not yet a solution. Their constructive theory rebuilt optimal control around finite *support* elements — small distinguished index sets that certify optimality, in the family spirit of [[George_Dantzig|Dantzig]]-style pivoting — yielding adaptive methods for linear and piecewise-linear problems, dual [[Feedback|feedback]] constructions, and algorithms that update the control as [[Sensor|sensor]] measurements arrive. The stance anticipates the logic of modern [[Model_predictive_control|model predictive control]] and [[Real-time_computing|real-time]] optimization: close the loop through repeated finite computations on [[Discrete_time_and_continuous_time|sampled]] data instead of storing an offline law for every state, with [[Linear_time-invariant_system|linear time-invariant]] structure exploited wherever it exists. Multi-volume treatises on constructive optimization methods followed through the 1980s and beyond, together with textbooks from which Belarusian and Soviet students learned [[Mathematical_optimization|mathematical programming]] and control. ## School, standing, and significance Kirillova spent her career at the Institute of Mathematics in Minsk, led its optimal-control research for decades, supervised a long line of candidates and doctors of science, and served on the editorial and program institutions of the discipline, including activity in the international federation for automatic control community. Her election as a corresponding member of the Belarusian academy made her one of the most senior women in Soviet-tradition [[Mathematics|mathematics]], in a specialty — [[Control_system|control]] — where women of her generation were rare; see [[List_of_systems_scientists]] for her place in the wider discipline. The through-line of her work is a [[Systems_engineering|systems engineer's]] ethic applied to pure theory: theorems about [[Optimal_control|optimality]] earn their keep when they terminate in an [[Algorithm|algorithm]] a [[Computer_science|computer]] can run before the plant drifts — a standard the whole field, from [[Robotics|robotics]] to [[Electrical_grid|grid]] dispatch, now takes for granted. **On the spine:** [[Optimal_control]] · [[Control_theory]] · [[Richard_Bellman]] · [[Cybernetics_in_the_Soviet_Union]]. ## Wikipedia : Wikitube **Strict pair:** [Wikipedia](https://en.wikipedia.org/wiki/Faina_Kirillova) : [Wikitube](https://en.wikitube.io/wiki/Faina_Kirillova) ## Previous hub tags Hubs: `Systems`. Portals: [[PORTAL_Systems_science]], [[PORTAL_Cybernetics]], [[PORTAL_Reliability_engineering]], [[PORTAL_Control_theory]], [[PORTAL_Systems_engineering]]. --- *Repopulated 2026-08-12 · redlink fill · 0 deletions.*