Differentiable Penalty Function Research Articles

Distributed Generalized Nash Equilibrium Seeking for Monotone Generalized Noncooperative Games by a Regularized Penalized Dynamical System

In this work, we study the generalized Nash equilibrium (GNE, see Definition 1) seeking problem for monotone generalized noncooperative games with set constraints and shared affine inequality constraints. A novel projected gradient-based regularized penalized dynamical system is proposed to solve this issue. The idea is to use a differentiable penalty function with a time-varying penalty parameter to deal with the inequality constraints. A time-varying regularization term is used to deal with the ill-poseness caused by the monotonicity assumption and the time-varying penalty term. The proposed dynamical system extends the regularized dynamical system in the literature to the projected gradient-based regularized penalized dynamical system, which can be used to solve generalized noncooperative games with set constraints and coupled constraints. Furthermore, we propose a distributed algorithm by using leader-following consensus, where the players have access to neighboring information only. For both cases, the asymptotic convergence to the least-norm variational equilibrium of the game is proven. Numerical examples show the effectiveness and efficiency of the proposed algorithms.

Defensive Ecological Adaptive Cruise Control Considering Neighboring Vehicles’ Blind-Spot Zones

This paper proposes a defensive ecological adaptive cruise control (DEco-ACC) algorithm that is capable of reducing an ego vehicle’s dwelling time in the blind spot zones (BSZs) of its neighboring vehicles. To this end, a model predictive control is applied in the use of information such as speed, position, and blind spot zones about preceding and neighboring vehicles. The cost function of the DEco-ACC consists of tracking performance, control effort, and dwelling time in BSZs. Specifically, a continuous and one-time differentiable penalty function is introduced to handle the constraints regarding the BSZs. For optimizing and evaluating the performance of the proposed DEco-ACC, real-world traffic data from Next Generation Simulation (NGSIM) are used to analyze and generate car-following scenarios during highway driving. Especially, in consideration of the most probable case that one neighboring vehicle exists at one adjacent lane, a parametric study is conducted to investigate the impact of the weighting factors on the performance of the DEco-ACC. The simulation results from 100 cases demonstrate that on average, the DEco-ACC with optimized weighting factors can reduce the dwelling time in the neighboring vehicles’ BSZs by 46.3% without significant deterioration of fuel consumption (0.04% increase in average fuel consumption) and drivability, as compared to the Eco-ACC, whose primary objective is the minimization of fuel consumption during safe car-following.

Nesterov Acceleration for Equality-Constrained Convex Optimization via Continuously Differentiable Penalty Functions

We propose a framework to use Nesterov's accelerated method for constrained convex optimization problems. Our approach consists of first reformulating the original problem as an unconstrained optimization problem using a continuously differentiable exact penalty function. This reformulation is based on replacing the Lagrange multipliers in the augmented Lagrangian of the original problem by Lagrange multiplier functions. The expressions of these Lagrange multiplier functions, which depend upon the gradients of the objective function and the constraints, can make the unconstrained penalty function non-convex in general even if the original problem is convex. We establish sufficient conditions on the objective function and the constraints of the original problem under which the unconstrained penalty function is convex. This enables us to use Nesterov's accelerated gradient method for unconstrained convex optimization and achieve a guaranteed rate of convergence which is better than the state-of-the-art first-order algorithms for constrained convex optimization. Simulations illustrate our results.

Sequential Stochastic Response Surface Method Using Moving Least Squares-Based Sparse Grid Scheme for Efficient Reliability Analysis

The present work demonstrates an efficient method for reliability analysis using sequential development of the stochastic response surface. Here, orthogonal Hermite polynomials are used whose unknown coefficients are evaluated using moving least square technique. To do so, collocation points in the conventional stochastic response surface method (SRSM) are replaced by the sparse grid scheme so as to reduce the number of function evaluations. Moreover, the domain is populated sequentially by the sparse grid based on the outcome of the optimization to find out the most probable failure point. Hence, the support points are generated based on a coupled effect of the optimization for failure region and the sub-grids hierarchy. Continuous and differentiable penalty function is imposed to determine multiple failure points, if any, by repeating the optimization. Once the response surface is developed, reliability analysis is carried out using importance sampling. Five different benchmark examples are presented in this study to validate the performance of the proposed modeling. As the accuracy of the method is established, two reliability-based design examples involving nonlinear finite element (FE) analysis of plates are demonstrated. Numerical study shows the efficiency of the proposed sequential SRSM in terms of accuracy and number of time-exhaustive evaluation of the original performance function, as compared to other methods available in the literature. Based on these results, it may be concluded that the proposed method works satisfactorily for a large class of reliability-based design problems.

On Smoothing l1 Exact Penalty Function for Constrained Optimization Problems

This article introduces a smoothing technique to the l1 exact penalty function. An application of the technique yields a twice continuously differentiable penalty function and a smoothed penalty problem. Under some mild conditions, the optimal solution to the smoothed penalty problem becomes an approximate optimal solution to the original constrained optimization problem. Based on the smoothed penalty problem, we propose an algorithm to solve the constrained optimization problem. Every limit point of the sequence generated by the algorithm is an optimal solution. Several numerical examples are presented to illustrate the performance of the proposed algorithm.

A Unified Approach to the Global Exactness of Penalty and Augmented Lagrangian Functions I: Parametric Exactness

In this two-part study we develop a unified approach to the analysis of the global exactness of various penalty and augmented Lagrangian functions for finite-dimensional constrained optimization problems. This approach allows one to verify in a simple and straightforward manner whether a given penalty/augmented Lagrangian function is exact, i.e. whether the problem of unconstrained minimization of this function is equivalent (in some sense) to the original constrained problem, provided the penalty parameter is sufficiently large. Our approach is based on the so-called localization principle that reduces the study of global exactness to a local analysis of a chosen merit function near globally optimal solutions. In turn, such local analysis can usually be performed with the use of sufficient optimality conditions and constraint qualifications. In the first paper we introduce the concept of global parametric exactness and derive the localization principle in the parametric form. With the use of this version of the localization principle we recover existing simple necessary and sufficient conditions for the global exactness of linear penalty functions, and for the existence of augmented Lagrange multipliers of Rockafellar-Wets' augmented Lagrangian. Also, we obtain completely new necessary and sufficient conditions for the global exactness of general nonlinear penalty functions, and for the global exactness of a continuously differentiable penalty function for nonlinear second-order cone programming problems. We briefly discuss how one can construct a continuously differentiable exact penalty function for nonlinear semidefinite programming problems, as well.

Proximal Limited-Memory Quasi-Newton Methods for Scenario-based Stochastic Optimal Control

Stochastic optimal control problems are typically of rather large scale involving millions of decision variables, but possess a certain structure which can be exploited by first-order methods such as forward-backward splitting and the alternating direction method of multipliers (ADMM). In this paper, we use the forward-backward envelope, a real-valued continuously differentiable penalty function, to recast the dual of the original nonsmooth problem as an unconstrained problem which we solve via the limited-memory BFGS algorithm. We show that the proposed method leads to a significant improvement of the convergence rate without increasing much the computational cost per iteration.

Nonlinear programming and Grossone: Quadratic Programing and the role of Constraint Qualifications

A novel and interesting approach to infinite and infinitesimal numbers was recently proposed in a series of papers and a book by Sergeyev. This novel numeral system is based on the use of a new infinite unit of measure (the number grossone, indicated by the numeral ①), the number of elements of the set, IN, of natural numbers. Based on the use of ①, De Cosmis and De Leone (2012) have then proposed a new exact differentiable penalty function for constrained optimization problems. In this paper these results are specialized to the important case of quadratic problems with linear constraints. Moreover, the crucial role of Constraint Qualification conditions (well know in constraint minimization literature) is also discussed with reference to the new proposed penalty function.

The Penalty Functions Method and Multiplier Rules Based on the Mordukhovich Subdifferential

We show that the finite-dimensional Fritz John multiplier rule, which is based on the limiting/Mordukhovich subdifferential, can be proved by using differentiable penalty functions and the basic calculus tools in variational analysis. The corresponding Kuhn–Tucker multiplier rule is derived from the Fritz John multiplier rule by imposing a constraint qualification condition or the exactness of an l1 penalty function. Complementing the existing proofs, our proofs provide another viewpoint on the fundamental multiplier rules employing the Mordukhovich subdifferential.

Scalable Nonlinear Programming via Exact Differentiable Penalty Functions and Trust-Region Newton Methods

We present an approach for nonlinear programming based on the direct minimization of an exact differentiable penalty function using trust-region Newton techniques. The approach provides desirable features required for scalability: it can detect and exploit directions of negative curvature, it is superlinearly convergent, and it enables the scalable computation of the Newton step through iterative linear algebra. Moreover, it presents features that are desirable for parametric optimization problems that must be solved in a latency-limited environment, as is the case for model predictive control and mixed-integer nonlinear programming. These features are fast detection of activity, efficient warm starting, and progress on a primal-dual merit function at every iteration. We note that other algorithmic approaches fail to satisfy at least one of these features. We derive general convergence results for our approach and demonstrate its behavior through numerical studies.

A simple smooth exact penalty function for smooth optimization problem

For smooth optimization problem with equality constraints, new continuously differentiable penalty function is derived. It is proved exact in the sense that local optimizers of a nonlinear program are precisely the optimizers of the associated penalty function under some nondegeneracy assumption. It is simple in the sense that the penalty function only includes the objective function and constrained functions, and it doesn’t include their gradients. This is achieved by augmenting the dimension of the program by a variable that controls the weight of the penalty terms.

The use of grossone in Mathematical Programming and Operations Research

The concepts of infinity and infinitesimal in mathematics date back to ancients Greek and have always attracted great attention. Very recently, a new methodology has been proposed by Sergeyev [10] for performing calculations with infinite and infinitesimal quantities, by introducing an infinite unit of measure expressed by the numeral ① (grossone). An important characteristic of this novel approach is its attention to numerical aspects. In this paper we will present some possible applications and use of ① in Operations Research and Mathematical Programming. In particular, we will show how the use of ① can be beneficial in anti-cycling procedure for the well-known Simplex Method for solving Linear Programming problems and in defining exact differentiable penalty functions in Nonlinear Programming.

Standard bi-quadratic optimization problems and unconstrained polynomial reformulations

A so-called Standard Bi-Quadratic Optimization Problem (StBQP) consists in minimizing a bi-quadratic form over the Cartesian product of two simplices (so this is different from a Bi-Standard QP where a quadratic function is minimized over the same set). An application example arises in portfolio selection. In this paper we present a bi-quartic formulation of StBQP, in order to get rid of the sign constraints. We study the first- and second-order optimality conditions of the original StBQP and the reformulated bi-quartic problem over the product of two Euclidean spheres. Furthermore, we discuss the one-to-one correspondence between the global/local solutions of StBQP and the global/local solutions of the reformulation. We introduce a continuously differentiable penalty function. Based upon this, the original problem is converted into the problem of locating an unconstrained global minimizer of a (specially structured) polynomial of degree eight.

Decrease of the Penalty Parameter in Differentiable Penalty Function Methods

We propose a simple modification to the differentiable penalty methods for solving nonlinear programming problems. This modification decreases the penalty parameter and the ill-conditioning of the penalty method and leads to a faster convergence to the optimal solution. We extend the modification to the augmented Lagrangian method and report some numerical results on several nonlinear programming test problems, showing the effectiveness of the proposed approach.

A New Exact Penalty Function

For constrained smooth or nonsmooth optimization problems, new continuously differentiable penalty functions are derived. They are proved exact in the sense that under some nondegeneracy assumption, local optimizers of a nonlinear program are precisely the optimizers of the associated penalty function. This is achieved by augmenting the dimension of the program by a variable that controls both the weight of the penalty terms and the regularization of the nonsmooth terms.

Structural design of doublet lenses with prespecified aberration targets

The authors present a systematic approach for the structural design of doublet lenses in accordance with a prespecified set of Gauss- ian parameters and primary aberration targets. This approach obviates the need for heuristic selection of glasses for the constituent lens ele- ments. An identical approach is implemented for the structural design of both cemented and broken contact doublets. The optimization procedure is the well-known least-squares method. For facilitating convergence a second-derivative damping technique is utilized. Constraints are taken care of by suitably defined continuously differentiable penalty functions that are treated as pseudoaberrations in optimization. A strategy for dy- namic weightage of the components of the defect function is devised for overcoming stagnation at the local optimum in the immediate neighbor- hood of the starting point. The strategy is also effective in constraining the glasses of the constituent elements of the doublet within the permis- sible glass domain. Some illustrative examples are given. © 1997 Society of Photo-Optical Instrumentation Engineers. (S0091-3286(97)02111-9)

ITERATIVE DYNAMIC PROGRAMMING FOR MINIMUM ENERGY CONTROL PROBLEMS WITH TIME DELAY

SUMMARYThis paper presents the use of iterative dynamic programming employing exact penalty functions for minimum energy control problems. We show that exact continuously non‐differentiable penalty functions are superior to continuously differentiable penalty functions in terms of satisfying final state constraints. We also demonstrate that the choice of an appropriate penalty function factor depends on the relative size of the time delay with respect to the final time and on the expected value of the energy consumption. A quadratic approximation (QA) of the delayed variables is much better than a linear approximation (LA) of the same for relatively large time delays. The QA improves the rate of convergence and avoids the formation of ‘kinks‘.A more general way of selecting appropriate penalty function factors is given and the results obtained using four illustrative examples of varying complexity corroborate the efficacy of the method.

Geometric approach to Fletcher's ideal penalty function

In this note, we derive a geometric formulation of an ideal penalty function for equality constrained problems. This differentiable penalty function requires no parameter estimation or adjustment, has numerical conditioning similar to that of the target function from which it is constructed, and also has the desirable property that the strict second-order constrained minima of the target function are precisely those strict second-order unconstrained minima of the penalty function which satisfy the constraints. Such a penalty function can be used to establish termination properties for algorithms which avoid ill-conditioned steps. Numerical values for the penalty function and its derivatives can be calculated efficiently using automatic differentiation techniques.

Equality and inequality constrained optimization algorithms with convergent stepsizes

In this paper, we consider two algorithms for nonlinear equality and inequality constrained optimization. Both algorithms utilize stepsize strategies based on differentiable penalty functions and quadratic programming subproblems. The essential difference between the algorithms is in the stepsize strategies used. The objective function in the quadratic subproblem includes a linear term that is dependent on the penalty functions. The quadratic objective function utilizes an approximate Hessian of the Lagrangian augmented by the penalty functions. In this approximation, it is possible to ignore the second-derivative terms arising from the constraints in the penalty functions.

Exact Penalty Functions in Constrained Optimization

In this paper formal definitions of exactness for penalty functions are introduced and sufficient conditions for a penalty function to be exact according to these definitions are stated, thus providing a unified framework for the study of both nondifferentiable and continuously differentiable penalty functions. In this framework the best-known classes of exact penalty functions are analyzed, and new results are established concerning the correspondence between the solutions of the constrained problem and the unconstrained minimizers of the penalty functions.