Finite Horizon Research Articles

Fast Nonlinear Model Predictive Control Combining Online Trajectory Optimization and Value Function Regression

ABSTRACTModel predictive control (MPC) is a well‐developed method capable of handling complex control tasks. Implementation of an MPC requires solution of a deterministic finite horizon nonlinear optimal control (OC) problem. OC problems can be solved globally, explicitly expressing the optimal policy (and value function) as a function of the present state, or, locally, using online trajectory optimization, generating a solution that is only valid for the present state. For nonlinear problems however, neither is possible analytically, and the latter, finding a local solution online, is usually preferred. The difficulty with online trajectory optimization is that the solution must be available within a single sampling period. The main parameter affecting the computational demand of MPC is the length of the prediction horizon. The goal in this work is to reduce the length of the prediction horizon of a model predictive controller (MPC) to reduce computation time whilst preserving optimality guarantees. To this end, we propose approximation of the trajectory optimization problem for MPC by learning a finite horizon value function. The approximated value function is inserted into a truncated trajectory optimization problem so that the MPC can be attained with a reduced prediction horizon, ergo reduced computational load. By sampling the value function in a goal oriented way, we show that an effective approximate value function can be found by including both the function value and the gradients of the value function. The result is an accurate approximate MPC which leverages learning methodologies to reduce computational cost while still accounting for constraints.

Random Walks and Lorentz Processes

Random walks and Lorentz processes serve as fundamental models for Brownian motion. The study of random walks is a favorite object of probability theory, whereas that of Lorentz processes belongs to the theory of hyperbolic dynamical systems. Here we first present an example where the method based on the probabilistic approach led to new results for the Lorentz process: concretely, the recurrence of the planar periodic Lorentz process with a finite horizon. Afterwards, an unsolved problem—related to a 1981 question of Sinai on locally perturbed periodic Lorentz processes—is formulated as an analogous problem in the language of random walks.

Linear State Optimal Control Problem with a Stochastic Switching Time

In this paper, we analyse an optimal control problem over a finite horizon with a stochastic switching time, assuming that the two optimal control problems present in its two stages have a particularly simple form called linear state. It is well known that linear state optimal control problems can be solved easily using the HJB equation approach and assuming that the value function is linear in the state. Unfortunately, this simplicity of solution does not extend to the problem with stochastic switching time. We prove that a necessary and sufficient condition for the problem to maintain a linear state structure is to assume that the hazard rate of the switching time depends only on the temporal variable. Finally, assuming that the hazard rate is constant, we completely characterise the solution of the obtained linear state optimal control problem.

Approximate solution of a zero-sum linear-quadratic differential game by the auxiliary parameter method: analytical/numerical study

ABSTRACT A zero-sum finite horizon linear-quadratic differential game is considered. First, the terminal-value problem for the game-theoretic Riccati matrix differential equation, associated with the considered game by the solvability conditions, is analysed. This problem may not have the solution in the entire time-interval of the game's duration. Using the method of auxiliary parameter, a sufficient condition (in the terms of the game's data) for the existence of the solution to the aforementioned terminal-value problem in the entire time-interval of the game's duration is derived. An approximate analytical solution to this problem also is obtained as a partial sum of some infinite series of functions arising in the method of auxiliary parameter. Based on this approximate solution, suboptimal state-feedback controls of the players are formally designed and justified. It is shown that the pair of these controls constitutes an approximate-saddle point in the considered game. The theoretical results of the article are illustrated by their application to approximate solution of the problem of pursuit-evasion engagement between two flying vehicles.

The efficiency of CUSUM schemes for monitoring the multivariate coefficient of variation in short runs process

Current monitoring technologies emphasize and address the issue of monitoring high-volume production processes. The high flexibility and diversity of current industrial production processes make monitoring technology for small batch processes even more important. In multivariate process monitoring, a broader applicability exists in multivariate coefficients of variation (MCV) based monitoring schemes due to the lower restriction of the process. In view of the effectiveness of MCV monitoring and with the aim to achieve further performance improvement of current MCV monitoring schemes in a finite horizon production, we additionally introduce two one-sided cumulative sum (CUSUM) MCV schemes. In the case of deterministic and random shifts, the design parameters of the proposed schemes are obtained via an optimization procedure designed by the Markov chain method and the corresponding performance is analysed based on different run length (RL) characteristics, including the mean and the standard deviation. Simulation comparisons with existing exponentially weighted moving average (EWMA) MCV schemes show that the proposed CUSUM MCV schemes are more efficient in monitoring most of the shifts, including the deterministic and random shifts. Finally, to demonstrate the benefits of the new monitoring schemes, a comprehensive case study on monitoring a steel sleeve manufacturing process is conducted.

Residence time density (RTd) identification through deconvolution of finite time horizon input output data and its Application to a novel analytical dispersion process RTd model

This work presents first an approach to the identification of Residence Time density (RTd) and Residence Time Distribution (RTD) functions, through deconvolution of finite time horizon, input/output data generated by a zero-order hold input, using the novel concept of zero-order hold, average, impulse response coefficients. Subsequently, a novel RTd analytical expression is derived for an open-open axial dispersion process with a finite experimental section in a tube of infinite length, and an explanation is provided for this RTd model’s differences from the literature. This new RTd expression is then employed in two case studies to demonstrate the effectiveness of the proposed identification through a deconvolution conceptual framework in identifying RTd model parameters, and in assisting in the selection of RTd models which are compatible with experimental data.

On the infinite time horizon approximation for Lévy-driven McKean-Vlasov SDEs with non-globally Lipschitz continuous and super-linearly growth drift and diffusion coefficients

This paper studies the numerical approximation for McKean-Vlasov stochastic differential equations driven by Lévy processes. We propose a tamed-adaptive Euler-Maruyama scheme and consider its strong convergence in both finite and infinite time horizons when applying for some classes of Lévy-driven McKean-Vlasov stochastic differential equations with non-globally Lipschitz continuous and super-linearly growth drift and diffusion coefficients.

Analysis of optimal portfolio on finite and small-time horizons for a stochastic volatility model with multiple correlated assets

Analysis of optimal portfolio on finite and small-time horizons for a stochastic volatility model with multiple correlated assets

A lie group PMP approach for optimal stabilization and tracking control of autonomous underwater vehicles

AbstractIn this research, we explore a finite horizon optimal stabilization and tracking control scheme for the dynamical model of a 6‐DOF Autonomous Underwater Vehicle (AUV). Dynamical equations of the AUV are represented in a Lie group () framework, encompassing both translational and rotational motions. Utilizing a left Lie group action on , we define error function for velocities via a right transport map to effectively address optimal trajectory tracking. The optimal control objective is formulated as a trade‐off problem, aiming to minimize both errors and control effort simultaneously. Left action on yields the left trivialized Hamiltonian function from which the concomitant state and costate dynamical equations are derived using Pontryagin's Minimum Principle (PMP). Consequently, the resulting two‐point boundary value problem is solved to obtain optimal trajectories. We demonstrate the optimality of the resulting solution obtained from the derived control law. For ensuring boundedness in the presence of small disturbances, this study incorporates the effects of internal parametric uncertainties associated with added mass and inertia components, along with the influence of external disturbances induced by ocean currents. Through simulation validations, we confirm the alignment of our results with the theoretical developments, demonstrating that the proposed control law effectively mitigates both parametric uncertainties and ocean current disturbances.

Portfolio problem for the α−hypergeometric stochastic volatility model with consumption

This paper solves the finite horizon optimal consumption-investment problem for an investor that trades in the α−hypergeometric stochastic volatility model, with preferences based on a power utility function. To find the optimal strategy, we follow the dynamic programming approach. First, we perform a suitable change of variables in order to fully transform the non-linear Hamilton-Jacobi-Bellman (HJB) equation into a semilinear one. Next, we apply the Girsanov theorem to deduce an implicit Feynman-Kac formula depending upon the volatility process. The operator defined by the Feynman-Kac representation is shown to be a contraction operator on a designed function space. Finally, the Banach fixed point theorem yields the existence of a solution to the HJB equation in the designed space. Moreover, we apply a verification theorem to guarantee that the solution of the HJB equation coincides with the value function.

Joint optimization of production, inspection, and maintenance under finite time for smart manufacturing systems

Given the flexible and configurable characteristics of smart manufacturing systems with a limited time per manufacturing task, the assumption of infinite time for prostems is no longer applicable to the joint-control strategy. Consequently, a joint-control model that considers production, inspection, and maintenance within a finite-time scenario for smart manufacturing systems is proposed in this paper. The objective is to optimize overall production and maintenance functions to minimize the total system cost. Comparing the joint strategy under infinite time with the proposed finite-time approach reveals significant differences in unit costs between the two scenarios. To enhance the effectiveness of the model, a discrete iterative algorithm with multiple loops was developed. Through a case study, it was observed that 1) joint strategies implemented within a finite time horizon were more cost-effective than those under infinite time, thus emphasizing the need for business managers to develop strategies within a finite time frame; 2) different production planning and efficiency levels had varying effects on the final joint strategy, necessitating customized strategies based on different production durations. Overall, the research gap regarding joint strategies within a finite-time context was addressed in this research, serving as a methodological foundation for practitioners to develop various strategies that minimize total costs across diverse real-world scenarios.

In defence of the desire for everlasting life: why secular faith cannot ground human meaning and solidarity

In this article, I argue that human meaning and value are grounded in an infinite horizon as opposed to the finite horizon of the building of a life. This infinite grounding of human meaning and value makes sense of and justifies the desire for everlasting life. I also argue that this infinite horizon can motivate an ethic of social justice better than the necessity of building a life within a finite timeframe could. In this article I take Martin Hägglund's This Life: Secular Faith and Spiritual Freedom as representative of the position that meaning and value can only be made sense of in light of the horizon of death; and I draw on phenomenologist Michel Henry's concept of Life and Jean‐Luc Marion's concept of the saturated phenomenon to argue against that position. I then draw on Friedrich Nietzsche, Karl Marx, and David Graeber to argue that social justice cannot be grounded in secular faith of temporal finitude but is rather best made sense of in view of an everlasting hope and a move to the infinite.

Robust Intelligent Nonlinear Predictive Control Based on Artificial Neural Network for Optimizing PMSM Drive Performance

In the realm of high-performance motor drive systems, achieving stable and optimal motor operation is crucial, particularly in environments characterized by disturbances. The implementation of model predictive control (MPC) represents a strategic methodology. However, conventional predictive controllers frequently encounter challenges due to uncertainties regarding the motor's internal parameters and external load characteristics, subsequently impacting the effectiveness of the control algorithm. This paper proposes a new robust predictive control combined with an artificial neural network (RMPC-ANN) approach applied to a permanent magnet synchronous motor (PMSM) to tackle challenges posed by external perturbations and parameter variations. The development of the proposed robust predictive controller involves optimizing a novel finite horizon cost function based on Taylor series expansion, which incorporates dual integral action into the control law. Crucially, this approach eliminates the necessity for measuring and observing external perturbations and parameter uncertainties. Additionally, for attaining high-precision speed control, the speed loop regulation relies on a multi-layer feedforward ANN algorithm. A comprehensive comparison was conducted using MATLAB/SIMULINK, assessing performance across diverse operating conditions. To further substantiate the numerical simulation results, a hardware-in-the-loop (HIL) configuration is implemented on the OPAL-RT platform, demonstrating the robustness and efficiency of the proposed control strategy.

Stages and main tasks of the century-long control theory and system identification development. Part VI. Model predictive control

The article is devoted to one of the relatively new areas in the field of control theory and practice, called model predictive control. The ideas underlying it, the connection with the theory of optimal control and numerical methods for solving optimization problems are described. It should be noted that currently the number of problems solved and control systems developed within the framework of this approach has come out on top in comparison with all others. This is evidenced by numerous publications concerning the theory and practical application of control methods and algorithms using a predictive model. The main distinctive feature of this method is that the control is not synthesized in advance, but is calculated directly during the operation of the system, i.e. at the current time. This made it possible, with such control, to implement original feedback that ensures stabilization of the process when the model is inaccurate and there is uncertainty associated with existing disturbances and measurements. It describes how such feedback is organized based on current measurement data. In this case, there is no need to interpret somehow the nature of the disturbances and errors acting on the system. Much attention in the article is paid to the possibilities that are admitted with this method of control. The model predictive control method seems especially attractive for systems with discrete impulse processes that are well described with using cognitive maps. It is considered an example of cognitive modeling of evolutionary demographic processes in Ukraine and the possibility of control them using a predictive model, when the model contains unstable eigenvalues. When solving problems of predictive control, it is proposed to use trajectory models, which directly follow from the Cauchy formula, in particular for discrete systems, instead of a local description in the state space. This makes it possible to take into account, when solving predictive control problems, how well or poorly controlled, observable or identifiable the system is. Non-standard capabilities of this method with a trajectory description appear during terminal control on a sliding interval with a finite horizon. A number of original formulations of terminal control problems are presented in the article and some of their properties are described.

An approach to mismatched disturbance rejection control for uncontrollable systems

This study focuses on the problem of optimal mismatched disturbance rejection control for uncontrollable linear discrete-time systems. In contrast to previous studies, by introducing a quadratic performance index such that the regulated state can track a reference trajectory and minimize the effects of disturbances, mismatched disturbance rejection control is transformed into a linear quadratic tracking problem. The necessary and sufficient conditions for the solvability of this problem over a finite horizon and a disturbance rejection controller are derived by solving a forward–backward difference equation. In the case of an infinite horizon, a sufficient condition for the stabilization of the system is obtained under the detectable condition. Additionally, in combination with the generalized extended state observer, a controller design method is proposed, and the stability analysis of the system under this controller is presented. This paper details our novel approach to disturbance rejection. Finally, four examples are provided to demonstrate the effectiveness of the proposed method.

Linear quadratic optimal control of stochastic 2-D Roesser models

This paper investigates the linear quadratic (LQ) optimal control problem for the stochastic two-dimensional (2-D) systems governed by Roesser models with multiplicative noise. The main contribution is to give the necessary and sufficient optimality condition by proposing a set of novel forward and backward stochastic partial difference equations (FBSPDE), and to further present the explicitly optimal feedback control laws on the finite horizon and on the infinite horizon based on the Riccati-like difference equations and the algebraic equation, respectively. Several numerical simulations are provided to illustrate the performance of the designed controllers.

Evaluation of counterparty credit risk under netting agreements

We investigate counterparty credit risk and credit valuation adjustments in portfolios including derivatives with early-exercise opportunities, under a netting agreement. We show that credit risk and netting agreements have a significant impact on the way portfolios are managed (that is, on options’ exercise strategies) and, therefore, on the value of the portfolio and on the price of counterparty risk. We derive the value of a netted portfolio as the solution of a zero-sum, finite horizon, discrete-time stochastic game. We show that this dynamic-game interpretation can be used to determine the value of the reglementary capital charges required of financial institutions to cover for counterparty credit risk and we propose a numerical valuation method. Numerical investigations show that currently used numerical approaches can grossly misestimate the value of credit valuation adjustments.

Optimal condition‐based parameter learning and mission abort decisions

AbstractUnexpected failures of safety‐critical systems during mission execution are not desirable in that they often result in severe safety hazards and significant financial losses. Prompt mission abort based on real‐time degradation data is an effective means to prevent such failures and enhance system safety. In this study, we focus on safety‐critical systems that experience cumulative shock degradation and fails when the degradation exceeds a failure threshold. Real‐time degradation measurements are obtained via sensor monitoring, which are stochastically related to the hidden degradation parameters that vary across components. We formulate the optimal mission risk control problem as a sequential abort decision‐making problem that integrates adaptive parameter learning, following which a dynamic Bayesian learning approach is exploited to sequentially infer the uncertain degradation parameters by utilizing real‐time sensor data. The problem is constituted as a finite horizon Markov decision process to minimize the expected costs associated with inspections, mission failures and system failures. We derive a series of structural properties of the value function and demonstrate the existence of optimal abort thresholds. In particular, we establish that the optimal policy follows a state‐dependent control limit policy. Additionally, we study the existence and monotonicity of control limits associated with both the number of inspections and degradation severities. We demonstrate the performance of the proposed risk management policy through comparative experiments that show substantial superiorities over risk‐induced loss control.

Conditional Lead-Time Flexibility in an Assemble-to-Order System

Problem definition: We introduce and formalize a concept termed conditional lead-time flexibility (CLF), which refers to the industry practice where a manufacturer requests that its upstream suppliers dynamically adjust the pipeline orders’ remaining lead times. Over a finite horizon, an assemble-to-order manufacturer makes joint decisions on inventory replenishment and lead-time adjustment to minimize the total discounted expected cost. Methodology/results: The problem is formulated as a multiperiod dynamic programming with an enlarged state space to track the adjustable pipeline orders. The replenishment and lead-time adjustment decisions are made sequentially in each period. The analysis reveals that (i) the optimal cost-to-go function is both convex and additively separable, that (ii) both the optimal replenishment and lead-time adjustment decisions follow general base-stock policies, and that (iii) the convexity and additive separability allow us to perform a derivative analysis on the optimality equations and design an efficient algorithm. Managerial implications: Using real industry data, we numerically find that (i) CLF achieves notable cost savings over a wide range of parameters, that (ii) the value of CLF comes from both the order expediting and deferring, and that (iii) CLF outperforms dual sourcing when the unit adjustment cost is less than 40% of the unit ordering cost from the express source. Our findings underscore the importance of CLF. Funding: T. Deng acknowledges financial support from the National Key Research and Development Program of China [Grant 2022YFB3305602] and the National Natural Science Foundation of China (NSFC) [Grant 72188101]. Supplemental Material: The electronic companion is available at https://doi.org/10.1287/msom.2023.0520 .

Event-triggered quasi-time-varying [formula omitted] filtering for switched systems via multiple trigger-dependent Lyapunov functionals

This paper investigates the event-triggered filtering problem for switched systems from the perspective of switched impulsive systems (SISs) by exploiting the characteristics of trigger and impulse behavior. By constructing multiple trigger-dependent Lyapunov functionals (MTDLFs), an event-triggered quasi-time-varying (ET-QTV) H∞ filtering scheme is developed. Since the MTDLFs properly capture the event-induced impulse property of SISs, they release less conservatism. The filtering stability in the scenario of frequent plant switching during an interevent interval is also guaranteed by incorporating the average dwell time switching. The filtering synthesis condition is formulated as a set of QTV matrix inequalities defined over a finite time horizon, based on which a sum of squares-based filter design algorithm is presented further. Additionally, the derivation of a lower bound on interevent intervals is clarified for the event-triggering mechanism (ETM) to exclude the Zeno phenomenon. Finally, the effectiveness and potential practicability of the proposed filtering scheme are validated using a morphing aircraft model.