IMT Institutional Repository: No conditions. Results ordered -Date Deposited. 2024-11-10T20:47:22ZEPrintshttp://eprints.imtlucca.it/images/logowhite.pnghttp://eprints.imtlucca.it/2018-01-16T10:14:31Z2018-01-16T10:14:31Zhttp://eprints.imtlucca.it/id/eprint/3863This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/38632018-01-16T10:14:31ZUncertainty-aware demand management of water distribution networks in deregulated energy marketsWe present an open-source solution for the operational control of drinking water distribution networks which accounts for the inherent uncertainty in water demand and electricity prices in the day-ahead market of a volatile deregulated economy. As increasingly more energy markets adopt this trading scheme, the operation of drinking water networks requires uncertainty-aware control approaches that mitigate the effect of volatility and result in an economic and safe operation of the network that meets the consumers’ need for uninterrupted water supply. We propose the use of scenario-based stochastic model predictive control: an advanced control methodology which comes at a considerable computation cost which is overcome by harnessing the parallelization capabilities of graphics processing units (GPUs) and using a massively parallelizable algorithm based on the accelerated proximal gradient method.Pantelis SopasakisAjay Kumar SampathiraoAlberto Bemporadalberto.bemporad@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.it2017-01-26T14:46:16Z2017-01-26T14:46:16Zhttp://eprints.imtlucca.it/id/eprint/3645This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/36452017-01-26T14:46:16ZStochastic gradient methods for stochastic model predictive controlWe introduce a new stochastic gradient algorithm, SAAGA, and investigate its employment for solving Stochastic MPC problems and multi-stage stochastic optimization programs in general. The method is particularly attractive for scenario-based formulations that involve a large number of scenarios, for which “batch” formulations may become inefficient due to high computational costs. Benefits of the method include cheap computations per iteration and fast convergence due to the sparsity of the proposed problem decomposition.A. ThemelisS. VillaPanagiotis PatrinosAlberto Bemporadalberto.bemporad@imtlucca.it2017-01-26T14:17:52Z2017-01-26T14:17:52Zhttp://eprints.imtlucca.it/id/eprint/3641This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/36412017-01-26T14:17:52ZGPU-accelerated stochastic predictive control of drinking water networksAjay Kumar SampathiraoPantelis SopasakisAlberto Bemporadalberto.bemporad@imtlucca.itPanagiotis Patrinos2016-10-10T15:08:08Z2016-10-10T15:08:08Zhttp://eprints.imtlucca.it/id/eprint/3583This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/35832016-10-10T15:08:08ZReal-time model predictive control based on dual gradient projection: Theory and fixed-point FPGA implementationThis paper proposes a method to design robust model predictive control (MPC) laws for discrete-time linear systems with hard mixed constraints on states and inputs, in case of only an inexact solution of the associated quadratic program is available, because of real-time requirements. By using a recently proposed dual gradient-projection algorithm, it is proved that the discrepancy of the optimal control law as compared with the obtained one is bounded even if the solver is implemented in fixed-point arithmetic. By defining an alternative MPC problem with tightened constraints, a feasible solution is obtained for the original MPC problem, which guarantees recursive feasibility and asymptotic stability of the closed-loop system with respect to a set including the origin, also considering the presence of external disturbances. The proposed MPC law is implemented on a field-programmable gate array in order to show the practical applicability of the method.Matteo RubagottiPanagiotis PatrinosAlberto GuiggianiAlberto Bemporadalberto.bemporad@imtlucca.it2015-12-11T11:32:13Z2015-12-11T11:32:13Zhttp://eprints.imtlucca.it/id/eprint/2971This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/29712015-12-11T11:32:13ZDouglas-rachford splitting: Complexity estimates and accelerated variantsWe propose a new approach for analyzing convergence of the Douglas-Rachford splitting method for solving convex composite optimization problems. The approach is based on a continuously differentiable function, the Douglas-Rachford Envelope (DRE), whose stationary points correspond to the solutions of the original (possibly nonsmooth) problem. By proving the equivalence between the Douglas-Rachford splitting method and a scaled gradient method applied to the DRE, results from smooth unconstrained optimization are employed to analyze convergence properties of DRS, to tune the method and to derive an accelerated version of it.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itLorenzo Stellalorenzo.stella@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2015-10-22T13:41:13Z2015-10-22T13:41:13Zhttp://eprints.imtlucca.it/id/eprint/2779This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/27792015-10-22T13:41:13ZDistributed solution of stochastic optimal control problems on GPUsStochastic optimal control problems arise in many applications and are, in principle, large-scale involving up to millions of decision variables. Their applicability in control applications is often limited by the availability of algorithms that can solve them efficiently and within the sampling time of the controlled system.
In this paper we propose a dual accelerated proximal gradient algorithm which is amenable to parallelization and demonstrate that its GPU implementation affords high speed-up values (with respect to a CPU implementation) and greatly outperforms well-established commercial optimizers such as Gurobi.Ajay Kumar SampathiraoPantelis Sopasakispantelis.sopasakis@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.it2015-05-19T09:35:54Z2015-10-28T14:47:41Zhttp://eprints.imtlucca.it/id/eprint/2681This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/26812015-05-19T09:35:54ZModel Predictive Control for Linear Impulsive Systems Linear impulsive control systems have been extensively studied with respect to their equilibrium points which, in most cases, are no other than the origin. However, the trajectory of an impulsive system cannot be stabilized to arbitrary desired points hindering their utilization in a great many applications. In this paper, we study the equilibrium of linear impulsive systems with respect to target-sets. We properly extend the notion of invariance and design stabilizing model predictive controllers (MPC). Finally, we apply the proposed methodology to control the intravenous bolus administration of Lithium.Pantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos SarimveisAlberto Bemporadalberto.bemporad@imtlucca.it2015-04-07T14:00:07Z2015-04-07T14:00:07Zhttp://eprints.imtlucca.it/id/eprint/2657This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/26572015-04-07T14:00:07ZA dual gradient-projection algorithm for model predictive control in fixed-point arithmetic Although linear Model Predictive Control has gained increasing popularity for controlling dynamical systems subject to constraints, the main barrier that prevents its widespread use in embedded applications is the need to solve a Quadratic Program (QP) in real-time. This paper proposes a dual gradient projection (DGP) algorithm specifically tailored for implementation on fixed-point hardware. A detailed convergence rate analysis is presented in the presence of round-off errors due to fixed-point arithmetic. Based on these results, concrete guidelines are provided for selecting the minimum number of fractional and integer bits that guarantee convergence to a suboptimal solution within a pre-specified tolerance, therefore reducing the cost and power consumption of the hardware device. Panagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Guiggianialberto.guiggiani@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2015-03-02T09:41:51Z2015-03-02T09:41:51Zhttp://eprints.imtlucca.it/id/eprint/2624This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/26242015-03-02T09:41:51ZA convex feasibility approach to anytime model predictive controlThis paper proposes to decouple performance optimization and enforcement of asymptotic convergence in Model Predictive Control (MPC) so that convergence to a given terminal set is achieved independently of how much performance is optimized at each sampling step. By embedding an explicit decreasing condition in the MPC constraints and thanks to a novel and very easy-to-implement convex feasibility solver proposed in the paper, it is possible to run an outer performance optimization algorithm on top of the feasibility solver and optimize for an amount of time that depends on the available CPU resources within the current sampling step (possibly going open-loop at a given sampling step in the extreme case no resources are available) and still guarantee convergence to the terminal set. While the MPC setup and the solver proposed in the paper can deal with quite general classes of functions, we highlight the synthesis method and show numerical results in case of linear MPC and ellipsoidal and polyhedral terminal sets. Alberto Bemporadalberto.bemporad@imtlucca.itDaniele Bernardinidaniele.bernardini@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.it2015-01-13T14:59:41Z2015-01-13T14:59:41Zhttp://eprints.imtlucca.it/id/eprint/2479This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/24792015-01-13T14:59:41ZA proximal alternating minimization method for L0-Regularized nonlinear optimization problems: application to state estimationIn this paper we consider the minimization of l0-regularized nonlinear optimization problems, where the objective function is the sum of a smooth convex term and the l0 quasi-norm of the decision variable. We introduce the class of coordinatewise minimizers and prove that any point in this class is a local minimum for our l0-regularized problem. Then, we devise a random proximal alternating minimization method, which has a simple iteration and is suitable for solving this class of optimization problems. Under convexity and coordinatewise Lipschitz gradient assumptions, we prove that any limit point of the sequence generated by our new algorithm belongs to the class of coordinatewise minimizers almost surely. We also show that the state estimation of dynamical systems with corrupted measurements can be modeled in our framework. Numerical experiments on state estimation of power systems, using IEEE bus test case, show that our algorithm performs favorably on solving such problemsAndrei - Mihai PatrascuIon NecoaraPanagiotis Patrinospanagiotis.patrinos@imtlucca.it2014-10-22T09:53:27Z2014-10-22T10:00:58Zhttp://eprints.imtlucca.it/id/eprint/2331This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/23312014-10-22T09:53:27ZStabilizing linear model predictive control under inexact numerical optimizationThis note describes a model predictive control (MPC) formulation for discrete-time linear systems with hard constraints on control and state variables, under the assumption that the solution of the associated quadratic program is neither optimal nor satisfies the inequality constraints. This is common in embedded control applications, for which real-time constraints and limited computing resources dictate restrictions on the possible number of on-line iterations that can be performed within a sampling period. The proposed approach is rather general, in that it does not refer to a particular optimization algorithm, and is based on the definition of an alternative MPC problem that we assume can only be solved within bounded levels of suboptimality, and violation of the inequality constraints. By showing that the inexact solution is a feasible suboptimal one for the original problem, asymptotic or exponential stability is guaranteed for the closed-loop system. Based on the above general results, we focus on a specific dual accelerated gradient-projection method to obtain a stabilizing MPC law that only requires a predetermined maximum number of on-line iterations.Matteo RubagottiPanagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-10-22T08:30:45Z2014-10-22T08:30:45Zhttp://eprints.imtlucca.it/id/eprint/2329This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/23292014-10-22T08:30:45ZMPC for power systems dispatch based on stochastic optimizationIn this paper we investigate the problem of optimal real-time power dispatch of an interconnection of conventional power generation plants, renewable resources and energy storage systems. The objective of the problem is to minimize imbalance costs and maximize the profit of the company managing the system whilst satisfying user demand. The managing company is able to trade energy on an electricity market. Energy prices on the market, user demand and intermittent generation from the renewable plants are considered stochastic processes. We show that under certain assumptions, the stochastic power dispatch problem over a finite horizon can be recast, under a proper choice for the feedback policies and for the disturbance set, into a stochastic optimization formulation but with deterministic constraints. We carry out a systematic study of stochastic optimization methods to solve this problem, in particular we analyze the stochastic gradient method. We also show that this problem can be approximated by a proper deterministic optimization problem using the sample average approximation method, which can then be solved by standard means.Ion NecoaraDragos Nicolae ClipiciPanagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-10-22T07:55:49Z2015-04-07T14:04:19Zhttp://eprints.imtlucca.it/id/eprint/2328This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/23282014-10-22T07:55:49ZFixed-point implementation of a proximal Newton method for embedded model predictive control (I)Extending the success of model predictive control (MPC) technologies in embedded applications heavily depends on the capability of improving quadratic programming (QP) solvers. Improvements can be done in two directions: better algorithms that reduce the number of arithmetic operations required to compute a solution, and more efficient architectures in terms of speed, power consumption, memory occupancy and cost. This paper proposes a fixed point implementation of a proximal Newton method to solve optimization problems arising in input-constrained MPC. The main advantages of the algorithm are its fast asymptotic convergence rate and its relatively low computational cost per iteration since it the solution of a small linear system is required. A detailed analysis on the effects of quantization errors is presented, showing the robustness of the algorithm with respect to finite-precision computations. A hardware implementation with specific optimizations to minimize computation times and memory footprint is also described, demonstrating the viability of low-cost, low-power controllers for high-bandwidth MPC applications. The algorithm is shown to be very effective for embedded MPC applications through a number of simulation experiments. Alberto Guiggianialberto.guiggiani@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-09-19T07:24:07Z2014-11-17T13:01:55Zhttp://eprints.imtlucca.it/id/eprint/2283This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/22832014-09-19T07:24:07ZStochastic model predictive control for constrained discrete-time Markovian switching systems In this paper we study constrained stochastic optimal control problems for Markovian switching systems, an extension of Markovian jump linear systems (MJLS), where the subsystems are allowed to be nonlinear. We develop appropriate notions of invariance and stability for such systems and provide terminal conditions for stochastic model predictive control (SMPC) that guarantee mean-square stability and robust constraint fulfillment of the Markovian switching system in closed-loop with the {SMPC} law under very weak assumptions. In the special but important case of constrained {MJLS} we present an algorithm for computing explicitly the {SMPC} control law off-line, that combines dynamic programming with parametric piecewise quadratic optimization. Panagiotis Patrinospanagiotis.patrinos@imtlucca.itPantelis Sopasakispantelis.sopasakis@imtlucca.itHaralambos SarimveisAlberto Bemporadalberto.bemporad@imtlucca.it2014-09-02T09:44:03Z2014-09-02T09:44:03Zhttp://eprints.imtlucca.it/id/eprint/2273This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/22732014-09-02T09:44:03ZDouglas-Rachford splitting: complexity estimates and accelerated variantsWe propose a new approach for analyzing convergence of the Douglas-Rachford splitting method for solving convex composite optimization problems. The approach is based on a continuously differentiable function, the Douglas-Rachford Envelope (DRE), whose stationary points correspond to the solutions of the original (possibly nonsmooth) problem. The Douglas-Rachford splitting method is shown to be equivalent to a scaled gradient method on the DRE, and so results from smooth unconstrained optimization are employed to analyze its convergence and optimally choose parameter {\gamma} and to derive an accelerated variant of Douglas-Rachford splitting. Panagiotis Patrinospanagiotis.patrinos@imtlucca.itLorenzo Stellalorenzo.stella@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-07-01T11:13:05Z2014-07-01T11:13:05Zhttp://eprints.imtlucca.it/id/eprint/2226This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/22262014-07-01T11:13:05ZProximal Newton methods for convex composite optimizationThis paper proposes two proximal Newton methods for convex nonsmooth optimization problems in composite form. The algorithms are based on a new continuously differentiable exact penalty function, namely the Composite Moreau Envelope. The first algorithm is based on a standard line search strategy, whereas the second one combines the global efficiency estimates of the corresponding first-order methods, while achieving fast asymptotic convergence rates. Furthermore, they are computationally attractive since each Newton iteration requires the solution of a linear system of usually small dimension.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-07-01T11:02:11Z2014-07-01T11:02:11Zhttp://eprints.imtlucca.it/id/eprint/2225This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/22252014-07-01T11:02:11ZForward-backward truncated Newton methods for convex composite optimizationThis paper proposes two proximal Newton-CG methods for convex nonsmooth optimization problems in composite form. The algorithms are based on a a reformulation of the original nonsmooth problem as the unconstrained minimization of a continuously differentiable function, namely the forward-backward envelope (FBE). The first algorithm is based on a standard line search strategy, whereas the second one combines the global efficiency estimates of the corresponding first-order methods, while achieving fast asymptotic convergence rates. Furthermore, they are computationally attractive since each Newton iteration requires the approximate solution of a linear system of usually small dimension. Panagiotis Patrinospanagiotis.patrinos@imtlucca.itLorenzo Stellalorenzo.stella@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-06-19T08:09:43Z2014-09-02T09:28:54Zhttp://eprints.imtlucca.it/id/eprint/2208This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/22082014-06-19T08:09:43ZRobust Model Predictive Control for optimal continuous drug administrationIn this paper the Model Predictive Control (MPC) technology is used for tackling the optimal drug administration problem. The important advantage of MPC compared to other control technologies is that it explicitly takes into account the constraints of the system. In particular,
for drug treatments of living organisms, MPC can guarantee satisfaction of the minimum toxic concentration (MTC) constraints. A whole-body physiologically-based pharmacokinetic (PBPK) model serves as the dynamic prediction model of the system after it is formulated as a
discrete-time state-space model. Only plasma measurements are assumed to be measured online. The rest of the states (drug concentrations in other organs and tissues) are estimated in real time by designing an artificial observer. The complete system (observer and MPC controller) is able to drive the drug concentration to the desired levels at the organs of interest, while satisfying the imposed constraints, even in the presence of modeling errors, disturbances and noise. A case study on a PBPK model with 7 compartments, constraints on 5 tissues and a variable drug concentration set-point illustrates the efficiency of the methodology in drug dosing control applications. The proposed methodology is also tested in an uncertain setting
and proves successful in presence of modelling errors and inaccurate measurements. Pantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2014-03-05T14:18:21Z2014-03-05T14:26:18Zhttp://eprints.imtlucca.it/id/eprint/2177This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/21772014-03-05T14:18:21ZReliability and efficiency for market parties in power systemsIn this paper we present control strategies for solving the problems of risk-averse bidding on the electricity markets, focusing on the Day-Ahead and Ancillary Services market, and of optimal real-time power dispatch from the point of view of a market participant, or Balance Responsible Party (BRP). For what concerns the bidding problem, the proposed algorithms are based on two-stage stochastic programming and are aimed at finding the optimal allocation of production between the day-ahead exchange market and the ancillary services market. For the real-time power dispatch problem, we devised a two-level hierarchical control strategy, where the upper-level computes economically optimal power set-points for the generators, and the lower level tracks them while considering constraints and dynamical models of the plant. Simulation results based on realistic data modeling the Dutch transmission network are shown to evaluate the effectiveness of the approach.Laura PugliaPanagiotis Patrinospanagiotis.patrinos@imtlucca.itDaniele Bernardinidaniele.bernardini@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-03-05T14:12:53Z2014-03-05T14:12:53Zhttp://eprints.imtlucca.it/id/eprint/2176This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/21762014-03-05T14:12:53ZStabilizing embedded MPC with computational complexity guaranteesThis paper describes a model predictive control (MPC) approach for discrete-time linear systems with hard constraints on control and state variables. The finite-horizon optimal control problem is formulated as a quadratic program (QP), and solved using a recently proposed dual fast gradient-projection method. More precisely, in a finite number of iterations of the mentioned optimization algorithm, a solution with bounded levels of infeasibility and suboptimality is determined for an alternative problem. This solution is shown to be a feasible suboptimal solution for the original problem, leading to exponential stability of the closed-loop system. The proposed strategy is particularly useful in embedded control applications, for which real-time constraints and limited computing resources can impose tight bounds on the possible number of iterations that can be performed within the scheduled sampling time.Matteo RubagottiPanagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-03-05T14:06:27Z2015-04-07T14:04:44Zhttp://eprints.imtlucca.it/id/eprint/2175This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/21752014-03-05T14:06:27ZFixed-point dual gradient projection for embedded model predictive controlAlthough linear Model Predictive Control has gained increasing popularity for controlling dynamical systems subject to constraints, the main barrier that prevents its widespread use in embedded applications is the need to solve a Quadratic Program (QP) in real-time. This paper proposes a dual gradient projection (DGP) algorithm specifically tailored for implementation on fixed-point hardware. A detailed convergence rate analysis is presented in the presence of round-off errors due to fixed-point arithmetic. Based on these results, concrete guidelines are provided for selecting the minimum number of fractional and integer bits that guarantee convergence to a suboptimal solution within a prespecified tolerance, therefore reducing the cost and power consumption of the hardware device.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Guiggianialberto.guiggiani@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2014-03-05T13:21:13Z2014-03-05T14:06:56Zhttp://eprints.imtlucca.it/id/eprint/2173This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/21732014-03-05T13:21:13ZAn accelerated dual gradient-projection algorithm for embedded linear model predictive controlThis paper proposes a dual fast gradient-projection method for solving quadratic programming problems that arise in model predictive control of linear systems subject to general polyhedral constraints on inputs and states. The proposed algorithm is well suited for embedded control applications in that: 1) it is extremely simple and easy to code; 2) the number of iterations to reach a given accuracy in terms of optimality and feasibility of the primal solution can be tightly estimated; and 3) the computational cost per iteration increases only linearly with the prediction horizon.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2013-10-25T08:30:43Z2014-06-16T10:17:42Zhttp://eprints.imtlucca.it/id/eprint/1841This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/18412013-10-25T08:30:43ZMPC for Sampled-Data Linear Systems: guaranteeing continuous-time positive invarianceModel Predictive Controllers (MPC) designed for sampled-data systems can be shown to violate the constraints in continuous time. A reformulation of the initial problem will guarantee constraint satisfaction throughout the intersample period. Polytopic inclusions of the continuous trajectory are used in this paper to establish additional constraints leading to a linearly constrained quadratic optimization problem. Continuous time asymptotic stability and continuous-time positive invariance are proven for the reformulated problem.Pantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2013-02-20T10:41:16Z2013-02-20T10:41:16Zhttp://eprints.imtlucca.it/id/eprint/1486This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/14862013-02-20T10:41:16ZSimple and Certifiable Quadratic Programming Algorithms for Embedded Linear Model Predictive ControlIn this paper we review a dual fast gradient-projection approach to solving quadratic programming (QP) problems recently proposed in [Patrinos and Bemporad, 2012] that is particularly useful for embedded model predictive control (MPC) of linear systems subject to linear constraints on inputs and states. We show that the method has a computational effort aligned with several other existing QP solvers typically used in MPC, and in addition it is extremely easy to code, requires only basic and easily parallelizable arithmetic operations, and a number of iterations to reach a given accuracy in terms of optimality and feasibility of the primal solution that can be estimated quite tightly by solving an off-line mixed-integer linear programming problem. This research was largely motivated by ongoing research activities on embedded MPC for aerospace systems carried out in collaboration with the European Space Agency.Alberto Bemporadalberto.bemporad@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.it2013-02-15T08:10:34Z2013-03-12T14:57:38Zhttp://eprints.imtlucca.it/id/eprint/1467This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/14672013-02-15T08:10:34ZDevelopment of a methodology for the computation of a near-optimal explicit control law for nonlinear systems that are subject to constraints coupling fuzzy model predictive control and multi-parametric programmingPantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2013-02-14T10:04:07Z2013-03-12T14:57:38Zhttp://eprints.imtlucca.it/id/eprint/1480This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/14802013-02-14T10:04:07ZModel Predictive Control for Linear Impulsive SystemsLinear Impulsive Control Systems have been extensively studied with respect to their equilibrium points which, in most cases, are no other than the origin. However, the trajectory of the system cannot be stabilized to arbitrary desired points which imposes a significant restriction towards their utilization in various applications such as drug administration. In this paper, we study the equilibrium of Linear Impulsive Systems in light of target-sets instead of the standard equilibrium point approach. We properly extend the notion of invariant sets which is crucial in designing asymptotically stable Model Predictive Controllers (MPC).Pantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos SarimveisAlberto Bemporadalberto.bemporad@imtlucca.it2013-02-12T12:03:40Z2013-02-12T12:03:40Zhttp://eprints.imtlucca.it/id/eprint/1470This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/14702013-02-12T12:03:40ZAn accelerated dual gradient-projection algorithm for linear model predictive controlThis paper proposes a dual fast gradient-projection method for solving quadratic programming problems that arise in linear model predictive control with general polyhedral constraints on inputs and states. The proposed algorithm is quite suitable for embedded control applications in that: (1) it is extremely simple and easy to code; (2) the number of iterations to reach a given accuracy in terms of optimality and feasibility of the primal solution can be estimated quite tightly; (3) the computational cost per iteration increases only linearly with the prediction horizon; and (4) the algorithm is also applicable to linear time-varying (LTV) model predictive control problems, with an extra on-line computational effort that is still linear with the prediction horizon.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itAlberto Bemporadalberto.bemporad@imtlucca.it2013-01-24T09:36:53Z2013-03-12T14:57:39Zhttp://eprints.imtlucca.it/id/eprint/1464This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/14642013-01-24T09:36:53ZDerivation of a feedback control law in explicit form for nonlinear systems coupling Fuzzy Model Predictive Control and multi-parametric programmingPantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2013-01-17T10:06:03Z2014-01-29T13:58:23Zhttp://eprints.imtlucca.it/id/eprint/1460This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/14602013-01-17T10:06:03ZTwo-time-scale MPC for economically optimal real-time operation of balance responsible partiesEuropean electrical networks are evolving towards a distributed system where the number of power plants is growing and also the green plants based on renewable energy sources (RES) like wind and solar are increasing. Integration of RES leads to energy imbalance, due to the difficulty to predict their production. This paper proposes a two-time-scale Hierarchical Model Predictive Control (HMPC) strategy for real-time optimal control of Balance Responsible Parties (BRPs) in power systems with high penetration of renewable energy sources (RES). The proposed control strategy is able to handle ramp-rate constraints efficiently and results in reduced generation and imbalance costs due to real-time economic optimization of power setpoints.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itDaniele Bernardinidaniele.bernardini@imtlucca.itAlessandro MaffeiAndrej JokicAlberto Bemporadalberto.bemporad@imtlucca.it2012-04-04T08:58:38Z2012-04-04T08:58:38Zhttp://eprints.imtlucca.it/id/eprint/1255This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/12552012-04-04T08:58:38ZStochastic MPC for real-time market-based optimal power dispatchWe formulate the problem of dynamic, real-time optimal power dispatch for electric power systems consisting of conventional power generators, intermittent generators from renewable sources, energy storage systems and price-inelastic loads. The generation company managing the power system can place bids on the real-time energy market (the so-called regulating market) in order to balance its loads and/or to make profit. Prices, demands and intermittent power injections are considered to be stochastic processes and the goal is to compute power injections for the conventional power generators, charge and discharge levels for the storage units and exchanged power with the rest of the grid that minimize operating and trading costs. We propose a scenario-based stochastic model predictive control algorithm to solve the real-time market-based optimal power dispatch problem.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itSergio TrimboliAlberto Bemporadalberto.bemporad@imtlucca.it2011-12-06T13:27:59Z2011-12-06T13:27:59Zhttp://eprints.imtlucca.it/id/eprint/1039This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10392011-12-06T13:27:59ZDevelopment of nonlinear quantitative structure-activity relationships using RBF networks and evolutionary computingQuantitative Structure Activity Relationships (QSARs) are mathematical models that correlate structural or property descriptions of compounds (hydrophobicity, topology, electronic properties etc.) with activities, such as chemical measurements and biological assays. In this paper we propose a modeling methodology suitable for QSAR studies which selects the proper descriptors based on evolutionary computing and finally produces Radial Basis Function (RBF) neural network models. The method is successfully applied to the benchmark Selwood data set.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itAlex AlexandridisAntreas AfantitisHaralambos SarimveisOlga Igglesi-Markopoulou2011-12-06T11:56:46Z2011-12-06T13:24:35Zhttp://eprints.imtlucca.it/id/eprint/1038This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10382011-12-06T11:56:46ZUsing the Radial Basis Function (RBF) neural network architecture to develop QSARs for the prediction of the toxicity of phenols in Tetrahymena pyriformisKalliopi MakridimaAntreas AfantitisGeorgia MelagrakiPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos SarimveisOlga Igglesi-Markopoulou2011-12-06T11:36:57Z2011-12-06T11:36:57Zhttp://eprints.imtlucca.it/id/eprint/1037This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10372011-12-06T11:36:57ZNeural network model-based paper machine marginal cost curvesIn this paper we present a methodology for developing paper machine marginal cost curves, which include variable material and energy costs, fixed costs and overhead costs. The methodology is based on the radial basis function (RBF) neural network architecture and takes into account the complex interactions between the different mill departments. The outcome of the proposed method is the calculation of marginal costs for different paper machine production rates and different grades. The resulting cost curves can be used to take optimal decisions regarding paper machine loadings and optimize production allocation.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos SarimveisTh RetsinaS.R RutherfordAlex Alexandridis2011-12-06T11:03:54Z2011-12-06T11:03:54Zhttp://eprints.imtlucca.it/id/eprint/1036This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10362011-12-06T11:03:54ZOptimal production scheduling for dairy industriesIn this work, a complete two-level framework for use in food and in particular dairy industries is proposed. The specific characteristics of the dairy industry have been taken into consideration, in terms of the behavior of food sales over time and the special requirements in the production phase. At the scheduling level, an MILP (Mixed Integer Linear Programming) model of the system was developed, using a continuous representation of time.Philip DoganisHaralambos SarimveisAlex AlexandridisPanagiotis Patrinospanagiotis.patrinos@imtlucca.it2011-12-06T10:51:35Z2011-12-06T10:51:35Zhttp://eprints.imtlucca.it/id/eprint/1035This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10352011-12-06T10:51:35ZAn RBF based neuro-dynamic approach for the control of stochastic dynamic systemsThis paper presents a neuro-dynamic programming methodology for the control of markov decision processes. The proposed method can be considered as a variant of the optimistic policy iteration, where radial basis function (RBF) networks are employed as a compact representation of the cost-to-go function and the ॕ-LSPE is used for policy evaluation. We also emphasize the reformulation of the Bellman equation around the post-decision state in order to circumvent the calculation of the expectation. The proposed algorithm is applied to a retailer-inventory management problem.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2011-12-06T10:39:31Z2011-12-06T10:39:31Zhttp://eprints.imtlucca.it/id/eprint/1034This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10342011-12-06T10:39:31ZAn explicit optimal control approach for mean-risk dynamic portfolio allocationPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2011-12-06T10:31:11Z2011-12-06T10:31:11Zhttp://eprints.imtlucca.it/id/eprint/1033This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10332011-12-06T10:31:11ZRobust optimal control: calculation of the explicit
control law combining dynamic programming and multiparametric optimizationPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2011-12-06T10:15:20Z2013-03-12T14:57:39Zhttp://eprints.imtlucca.it/id/eprint/1032This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10322011-12-06T10:15:20ZExplicit control for nonlinear constrained systems combining fuzzy model predictive control and multiparametric
programmingPantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2011-12-05T15:49:54Z2013-03-12T14:57:38Zhttp://eprints.imtlucca.it/id/eprint/1031This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10312011-12-05T15:49:54ZPhysiologically based pharmacokinetic modeling and predictive control: an integrated approach for optimal drug administrationThe barriers between systems engineering and medicine are slowly eroding as recently it has become evident that medicine has a lot to gain from systems technology. In particular, the drug administration problem can be cast as a control engineering problem, where the objective is to keep the drug concentration at certain organs in the body close to desired set-points. A number of constraints render the problem rather challenging. For example, hard constraints may be posed on drug concentration, because if it exceeds an upper limit, the effects of the drug are adverse and toxic.
In this paper we show that a popular method in control engineering can be used for determining the optimal drug administration. Specifically, the Model Predictive Control (MPC) technology can be adopted for taking optimal decisions regarding regulation of drug concentration in the human body, while posing constraints on both drug concentration and drug infusion rate.Pantelis Sopasakispantelis.sopasakis@imtlucca.itPanagiotis Patrinospanagiotis.patrinos@imtlucca.itStefania GiannikouHaralambos Sarimveis2011-12-05T14:19:43Z2013-03-12T14:57:38Zhttp://eprints.imtlucca.it/id/eprint/1030This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10302011-12-05T14:19:43ZStochastic model predictive control for constrained networked control systems with random time delayIn this paper the continuous time stochastic constrained optimal control problem is formulated for the class of networked control systems assuming that time delays follow a discrete-time, finite Markov chain . Polytopic overapproximations of the system's trajectories are employed to produce a polyhedral inner approximation of the non-convex constraint set resulting from imposing the constraints in continuous time. The problem is cast in a Markov jump linear systems (MJLS) framework and a stochastic MPC controller is calculated explicitly, oine, coupling dynamic programming with parametric piecewise quadratic (PWQ) optimization. The calculated control law leads to stochastic stability of the closed loop system, in the mean square sense and respects the state and input constraints in continuous time.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itPantelis Sopasakispantelis.sopasakis@imtlucca.itHaralambos Sarimveis2011-12-05T11:40:26Z2011-12-05T11:40:26Zhttp://eprints.imtlucca.it/id/eprint/1027This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10272011-12-05T11:40:26ZA two-stage evolutionary algorithm for variable selection in the development of RBF neural network modelsIn many modeling problems that are based on input–output data, information about a plethora of variables is available. In these cases, the proper selection of explanatory variables is very critical for the success of the produced model, since it eliminates noisy variables and possible correlations, reduces the size of the model and accomplishes more accurate predictions. Many variable selection procedures have been proposed in the literature, but most of them consider only linear models. In this work, we present a novel methodology for variable selection in nonlinear modeling, which combines the advantages of several artificial intelligence technologies. More specifically, the Radial Basis Function (RBF) neural network architecture serves as the nonlinear modeling tool, by exploiting the simplicity of its topology and the fast fuzzy means training algorithm. The proper variables are selected in two stages using a multi-objective optimization approach: in the first stage, a specially designed genetic algorithm minimizes the prediction error over a monitoring data set, while in the second stage a simulated annealing technique aims at the reduction of the number of explanatory variables. The efficiency of the proposed method is illustrated through its application to a number of benchmark problems.Alex AlexandridisPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos SarimveisGeorge Tsekouras2011-12-05T11:15:09Z2011-12-05T11:15:09Zhttp://eprints.imtlucca.it/id/eprint/1026This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10262011-12-05T11:15:09ZTime series sales forecasting for short shelf-life food products based on artificial neural networks and evolutionary computingDue to the strong competition that exists today, most manufacturing organizations are in a continuous effort for increasing their profits and reducing their costs. Accurate sales forecasting is certainly an inexpensive way to meet the aforementioned goals, since this leads to improved customer service, reduced lost sales and product returns and more efficient production planning. Especially for the food industry, successful sales forecasting systems can be very beneficial, due to the short shelf-life of many food products and the importance of the product quality which is closely related to human health. In this paper we present a complete framework that can be used for developing nonlinear time series sales forecasting models. The method is a combination of two artificial intelligence technologies, namely the radial basis function (RBF) neural network architecture and a specially designed genetic algorithm (GA). The methodology is applied successfully to sales data of fresh milk provided by a major manufacturing company of dairy products.Philip DoganisAlex AlexandridisPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2011-12-05T11:00:57Z2011-12-05T11:00:57Zhttp://eprints.imtlucca.it/id/eprint/1025This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10252011-12-05T11:00:57ZDynamic modeling and control of supply chain systems: a reviewSupply chains are complicated dynamical systems triggered by customer demands. Proper selection of equipment, machinery, buildings and transportation fleets is a key component for the success of such systems. However, efficiency of supply chains mostly depends on management decisions, which are often based on intuition and experience. Due to the increasing complexity of supply chain systems (which is the result of changes in customer preferences, the globalization of the economy and the stringy competition among companies), these decisions are often far from optimum. Another factor that causes difficulties in decision making is that different stages in supply chains are often supervised by different groups of people with different managing philosophies. From the early 1950s it became evident that a rigorous framework for analyzing the dynamics of supply chains and taking proper decisions could improve substantially the performance of the systems. Due to the resemblance of supply chains to engineering dynamical systems, control theory has provided a solid background for building such a framework. During the last half century many mathematical tools emerging from the control literature have been applied to the supply chain management problem. These tools vary from classical transfer function analysis to highly sophisticated control methodologies, such as model predictive control (MPC) and neuro-dynamic programming. The aim of this paper is to provide a review of this effort. The reader will find representative references of many alternative control philosophies and identify the advantages, weaknesses and complexities of each one. The bottom line of this review is that a joint co-operation between control experts and supply chain managers has the potential to introduce more realism to the dynamical models and develop improved supply chain management policies.Haralambos SarimveisPanagiotis Patrinospanagiotis.patrinos@imtlucca.itChris D. TarantilisChris T. Kiranoudis2011-12-05T10:30:11Z2011-12-05T10:30:11Zhttp://eprints.imtlucca.it/id/eprint/1024This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10242011-12-05T10:30:11ZVariable selection in nonlinear modeling based on RBF networks and evolutionary computationIn this paper a novel variable selection method based on Radial Basis Function (RBF) neural networks and genetic algorithms is presented. The fuzzy means algorithm is utilized as the training method for the RBF networks, due to its inherent speed, the deterministic approach of selecting the hidden node centers and the fact that it involves only a single tuning parameter. The trade-off between the accuracy and parsimony of the produced model is handled by using Final Prediction Error criterion, based on the RBF training and validation errors, as a fitness function of the proposed genetic algorithm. The tuning parameter required by the fuzzy means algorithm is treated as a free variable by the genetic algorithm. The proposed method was tested in benchmark data sets stemming from the scientific communities of time-series prediction and medicinal chemistry and produced promising results.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itAlex AlexandridisKonstantinos NinosHaralambos Sarimveis2011-12-05T10:20:42Z2011-12-05T10:20:42Zhttp://eprints.imtlucca.it/id/eprint/1023This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10232011-12-05T10:20:42ZA new algorithm for solving convex parametric quadratic programs based on graphical derivatives of solution mappingsIn this paper we derive formulas for computing graphical derivatives of the (possibly multivalued) solution mapping for convex parametric quadratic programs. Parametric programming has recently received much attention in the control community, however most algorithms are based on the restrictive assumption that the so called critical regions of the solution form a polyhedral subdivision, i.e. the intersection of two critical regions is either empty or a face of both regions. Based on the theoretical results of this paper, we relax this assumption and show how we can efficiently compute all adjacent full dimensional critical regions along a facet of an already discovered critical region. Coupling the proposed approach with the graph traversal paradigm, we obtain very efficient algorithms for the solution of parametric convex quadratic programsPanagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis2011-12-05T09:47:20Z2013-03-12T14:57:38Zhttp://eprints.imtlucca.it/id/eprint/1022This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10222011-12-05T09:47:20ZA global piecewise smooth Newton method for fast large-scale model predictive controlIn this paper, the strictly convex quadratic program (QP) arising in model predictive control (MPC) for constrained linear systems is reformulated as a system of piecewise affine equations. A regularized piecewise smooth Newton method with exact line search on a convex, differentiable, piecewise-quadratic merit function is proposed for the solution of the reformulated problem. The algorithm has considerable merits when applied to MPC over standard active set or interior point algorithms. Its performance is tested and compared against state-of-the-art QP solvers on a series of benchmark problems. The proposed algorithm is orders of magnitudes faster, especially for large-scale problems and long horizons. For example, for the challenging crude distillation unit model of Pannocchia, Rawlings, and Wright (2007) with 252 states, 32 inputs, and 90 outputs, the average running time of the proposed approach is 1.57 ms.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos SarimveisPantelis Sopasakispantelis.sopasakis@imtlucca.it2011-12-05T09:38:04Z2011-12-05T09:38:04Zhttp://eprints.imtlucca.it/id/eprint/1021This item is in the repository with the URL: http://eprints.imtlucca.it/id/eprint/10212011-12-05T09:38:04ZConvex parametric piecewise quadratic optimization: theory and algorithmsIn this paper we study the problem of parametric minimization of convex piecewise quadratic functions. Our study provides a unifying framework for convex parametric quadratic and linear programs. Furthermore, it extends parametric optimization algorithms to problems with piecewise quadratic cost functions, paving the way for new applications of parametric optimization in explicit dynamic programming and optimal control with quadratic stage cost.Panagiotis Patrinospanagiotis.patrinos@imtlucca.itHaralambos Sarimveis