A worst disturbance design criterion in the theory of analytical control systems synthesis
Analytical design techniques of linear optimal control applied to minimax problem where disturbance energy is bounded
SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Analytical design techniques of linear optimal control applied to minimax problem where disturbance energy is bounded
Consideration is given to classical and minimax problems involved in aeroassisted transfer from high earth orbit (HEO) to low earth orbit (LEO). The transfer is restricted to coplanar operation, with trajectory control effected by means of lift modulation. The performance of the maneuver is indexed to the energy expenditure or, alternatively, the time integral of the heating rate. Firist-order optimality conditions are defined for the classical approach, as are a sequential gradient-restoration algorithm and a combined gradient-restoration algorithm. Minimization techniques are presented for the aeroassisted transfer energy consumption and time-delay integral of the heating rate, as well as minimization of the pressure. It is shown that the eigenvalues of the Jacobian matrix of the differential system is both stiff and unstable, implying that the sequential gradient restoration algorithm in its present version is unsuitable. A new method, involving a multipoint approach to the two-poing boundary value problem, is recommended.
Both classical and minimax problems of optimal control arising in the study of noncoplanar, aeroassisted orbital transfer are considered and are illustrated with the example of LEO-to-LEO transfer. Trajectory control is achieved by modulation of the lift coefficient and the angle of bank. Problems considered include the minimization of the energy required for orbital transfer, maximization of the flight time during the atmospheric portion of the trajectory, and minimization of the peak heating rate. The near-grazing solution is found to be a good compromise between energy and heating requirements.
Necessary conditions for minimax problems with isolated or flat maxima are presented. Some relevant properties concerning the peak heating rate and the peak deceleration during atmospheric entry are discussed. As application of the theory, the problem of minimizing the peak heating rate of a skip trajectory is solved with special emphasis on the discussion of the continuity of the lift control at the point where the maximum occurs along the trajectory.
Chebyshev minimax solutions for Saturn launch vehicle control problems, and hybrid-analog computer realization of algorithm
The maneuver considered in the present investigation involves the coplanar transfer of a spacecraft from a high earth orbit (HEO) to a low earth orbit (LEO). HEO can be a geosynchronous earth orbit (GEO). The basic concept utilized involves the hybrid combination of propulsive maneuvers in space and aerodynamic maneuvers in the sensible atmosphere. The considered type of flight is also called synergetic space flight. With respect to the atmospheric part of the maneuver, trajectory control is achieved by means of lift modulation. The Bolza problem of optimal control is stated, and the first-order optimality conditions for this problem are given. The one-arc approach, the two-arc approach, and the three-subarc approach are discussed. Attention is given to the Chebyshev problem of optimal control, details concerning aeroassisted orbital transfer (AOT), AOT optimization problems, and numerical experiments.
The optimal-control problem of abort-landing trajectories in the presence of low-altitude wind shear is investigated analytically. The vertical-plane Newtonian motion of a point-mass aircraft in a steady wind field is modeled, and a sequential gradient-restoration algorithm is applied. Numerical results showing the effects of wind-shear intensity, initial altitude, and power-setting rate are presented in extensive graphs and discussed in detail. Optimal trajectories for strong or severe wind shears are found to begin with a descent, followed by level flight and then an ascent after leaving the shear region.
This paper presents the necessary conditions for solving Chebyshev minimax (or maximax) problems with bounded control. The jump conditions obtained are applicable to problems with single or multiple maxima. By using Contensou domain of maneuverability, it is shown that when the maxima are isolated single points the control is generally continuous at the jump point in the minimax problems and discontinuous in the maximax problems in which the first time derivative of the maximax function contains the control variable. The theory is applied to the problem of maximizing the flight radius in a closed circuit glide of a hypervelocity vehicle and to a maximax optimal control problem in which the control appears explicitly with the first time derivative of the maximax function.
The abort landing problem is considered with reference to flight in a vertical plane. It is assumed that, upon sensing that the aircraft is in a windshear, the pilot increases the power setting at a constant time rate until maximum power setting is reached; afterward, the power setting is held constant. The performance index being minimized is the peak value of the altitude drop; the resulting optimization problem is a minimax or Chebyshev problem of optimal control. It is found that, for strong-to-severe windshears, the optimal trajectory includes three branches: a descending flight branch followed by a nearly horizontal flight branch, followed by an ascending flight branch after the aircraft has passed through the shear region. The peak altitude drop depends on the windshear intensity, the initial altitude, and the power setting rate; it increases as the windshear intensity increases and the initial altitude increases, and it decreases as the power setting rate increases.
The problem of optimal feedback control of uncertain discrete-time dynamic systems is considered where the uncertain quantities do not have a stochastic description but instead are known to belong to given sets. The problem is converted to a sequential minimax problem and dynamic programming is suggested as a general method for its solution. The notion of a sufficiently informative function, which parallels the notion of a sufficient statistic of stochastic optimal control, is introduced, and conditions under which the optimal controller decomposes into an estimator and an actuator are identified.
In this paper the problem of optimal feedback control of uncertain discrete-time dynamic systems is considered where the uncertain quantities do not have a stochastic description but instead are known to belong to given sets. The problem is converted to a sequential minimax problem and dynamic programming is suggested as a general method for its solution. The notion of a sufficiently informative function, which parallels the notion of a sufficient statistic of stochastic optimal control, is introduced, and conditions under which the optimal controller decomposes into an estimator and an actuator are identified.
The operation of some networks, such as air transportation networks, can be complicated by congestion through a small subset of nodes. The congestion may be influenced by the connectivity of the network, or by the presence of constraints restricting the flow through particular nodes. This work investigates the effects of both connectivity and node flow constraints on the operation of a network. We develop the Minimax Node Load Problem (MNLP), a multicommodity flow model which minimizes the worst-case flow through any node in a given input network. The optimal solution to this problem provides us with the minimax node load, which we propose as a measure of network congestion. Keeping the number of nodes fixed, we first increase connectivity in a series of networks, and observe that topologies with more distributed connections result in a reduction in the minimax node load. However, when connectivity is increased further, the reductions diminish and are accompanied by solutions with undesirable qualities such as longer commodity paths. We then perform a second set of experiments over the same network, constraining flow through different subsets of nodes at different magnitudes of flow restriction, finding that (1) more constrained nodes lead to the largest increases in minimax node load and (2) constraints on the most connected nodes have the greatest effect on both congestion and commodity path length.
The operation of some networks, such as air transportation networks, can be complicated by congestion through a small subset of nodes. The congestion may be influenced by the connectivity of the network, or by the presence of constraints restricting the flow through particular nodes. This work investigates the effects of both connectivity and node flow constraints on the operation of a network. We develop the Minimax Node Load Problem (MNLP), a multicommodity flow model which minimizes the worst-case flow through any node in a given input network. The optimal solution to this problem provides us with the minimax node load, which we propose as a measure of network congestion. Keeping the number of nodes fixed, we first increase connectivity in a series of networks, and observe that topologies with more distributed connections result in a reduction in the minimax node load. However, when connectivity is increased further, the reductions diminish and are accompanied by solutions with undesirable qualities such as longer commodity paths. We then perform a second set of experiments over the same network, constraining flow through different subsets of nodes at different magnitudes of flow restriction, finding that (1) more constrained nodes lead to the largest increases in minimax node load and (2) constraints on the most connected nodes have the greatest effect on both congestion and commodity path length.
For a nonlinear system of differential equations with aftereffect, two mutually complement game minimax (maximin) problems for the quality functional are considered. Assuming that a part of phase coordinates of the system is measured (with error) sufficiently frequently, we provide solving algorithms that are stable with respect to the information noise and computational errors. The proposed algorithms are based on the Krasovskii extremal translation principle.
Earlier, we proved a duality between two optimizations problems [Phys. Rev. Lett. 104, 120501 (2010)]. The primary one is, given two quantum channels M and N, to find a quantum channel R such that R White-Bullet N is optimally close to M as measured by the worst-case entanglement fidelity. The dual problem involves the information obtained by the environment through the so-called complementary channels M and N, and consists in finding a quantum channel R' such that R Prime White-Bullet cM is optimally close to N. It turns out to be easier to find an approximate solution to the dual problem in certain important situations, notably when M is the identity channel - the problem of quantum error correction - yielding a good near-optimal worst-case entanglement fidelity as well as the corresponding near-optimal correcting channel. Here we provide more detailed proofs of these results. In addition, we generalize the main theorem to the case where there are certain constraints on the implementation of R, namely, on the number of Kraus operators. We also offer a simple algebraic form for the near-optimal correction channel in the case M=id. For approximate error correction, we show that any {epsilon}-correctable channel is, up to appending an ancilla, {epsilon}-close to an exactly correctable one. We also demonstrate an application of our theorem to the problem of minimax state discrimination.
This final technical report summarizes the outcomes of a DOE-funded project on federated scientific machine learning (FL) under nonsmooth, nonconvex, and hierarchical optimization settings. The project develops new mathematical models, algorithms, and theoretical guarantees for decentralized stochastic, bilevel, and minimax optimization problems arising in DOE mission-relevant applications. A unified framework of randomized and zeroth-order federated optimization methods is introduced, providing provable convergence, communication efficiency, and sample-complexity guarantees. The report documents algorithmic design, theoretical analysis, and empirical validation of the proposed federated learning methods. The project also contributes to workforce development through graduate training and dissemination of results via publications and seminars.
General, closed-form, analytical solutions are determined for certain classes of C-minimax control problems, several alternative mathematical theories are derived, and a controller design theory is developed to give optimal control in the presence of unmeasureable external disturbances.
The Failure Detection and Identification (FDI) process is viewed as consisting of two stages: residual generation and decision making. It is argued that a robust FDI system can be achieved by designing a robust residual generation process. Analytical redundancy, the basis for residual generation, is characterized in terms of a parity space. Using the concept of parity relations, residuals can be generated in a number of ways and the design of a robust residual generation process can be formulated as a minimax optimization problem. An example is included to illustrate this design methodology. Previously announcedd in STAR as N83-20653