Nonconvex Hamiltonians Research Articles

A comparison principle for semilinear Hamilton–Jacobi–Bellman equations in the Wasserstein space

The goal of this paper is to prove a comparison principle for viscosity solutions of semilinear Hamilton–Jacobi equations in the space of probability measures. The method involves leveraging differentiability properties of the 2-Wasserstein distance in the doubling of variables argument, which is done by introducing a further entropy penalization that ensures that the relevant optima are achieved at positive, Lipschitz continuous densities with finite Fischer information. This allows to prove uniqueness and stability of viscosity solutions in the class of bounded Lipschitz continuous (with respect to the 1-Wasserstein distance) functions. The result does not appeal to a mean field control formulation of the equation, and, as such, applies to equations with nonconvex Hamiltonians and measure-dependent volatility. For convex Hamiltonians that derive from a potential, we prove that the value function associated with a suitable mean-field optimal control problem with nondegenerate idiosyncratic noise is indeed the unique viscosity solution.

Slow periodic homogenization for Hamilton–jacobi equations

Capuzzo-Dolcetta–Ishii proved that the rate of periodic homogenization for coercive Hamilton–Jacobi equations is . We complement this result by constructing examples of coercive nonconvex Hamiltonians whose rate of periodic homogenization is .

The vanishing discount problem for monotone systems of Hamilton-Jacobi equations: a counterexample to the full convergence

<abstract><p>In recent years there has been intense interest in the vanishing discount problem for Hamilton-Jacobi equations. In the case of the scalar equation, B. Ziliotto has recently given an example of the Hamilton-Jacobi equation having non-convex Hamiltonian in the gradient variable, for which the full convergence of the solutions does not hold as the discount factor tends to zero. We give here an explicit example of nonlinear monotone systems of Hamilton-Jacobi equations having convex Hamiltonians in the gradient variable, for which the full convergence of the solutions fails as the discount factor goes to zero.</p></abstract>

Hydrodynamic limit for a 2D interlaced particle process

The Markov dynamics of interlaced particle arrays, introduced by A. Borodin and P. L. Ferrari in (Comm. Math. Phys. 325 (2014) 603–684), is a classical example of (2+1)-dimensional random growth model belonging to the so-called Anisotropic KPZ universality class. In (Comm. Pure Appl. Math. 72 (2018) 620–666), a hydrodynamic limit—the convergence of the height profile, after space/time rescaling, to the solution of a deterministic Hamilton–Jacobi PDE with nonconvex Hamiltonian—was proven when either the initial profile is convex, or for small times, before the solution develops shocks. In the present work, we give a simpler proof, that works for all times and for all initial profiles for which the limit equation makes sense. In particular, the convexity assumption is dropped. The main new idea is a new viewpoint about “finite speed of propagation” that allows to bypass the need of a priori control of the interface gradients, or equivalently of inter-particle distances.

An example of failure of stochastic homogenization for viscous Hamilton-Jacobi equations without convexity

We give an example of the failure of homogenization for a viscous Hamilton-Jacobi equation with non-convex Hamiltonian.

A Kernel-Based explicit unconditionally stable scheme for Hamilton-Jacobi equations on nonuniform meshes

In [11], the authors developed a class of high-order numerical schemes for the Hamilton-Jacobi (H-J) equations, which are unconditionally stable, yet take the form of an explicit scheme. This paper extends such schemes, so that they are more effective at capturing sharp gradients, especially on nonuniform meshes. In particular, we modify the weighted essentially non-oscillatory (WENO) methodology in the previously developed schemes by incorporating an exponential basis and adapting the previously developed nonlinear filters used to control oscillations. The main advantages of the proposed schemes are their effectiveness and simplicity, since they can be easily implemented on higher-dimensional nonuniform meshes. We perform numerical experiments on a collection of examples, including H-J equations with linear, nonlinear, convex and non-convex Hamiltonians. To demonstrate the flexibility of the proposed schemes, we also include test problems defined on non-trivial geometry.

The Level-Set Method for Multi-Material Wet Etching and Non-Planar Selective Epitaxy

We present numerical methods to enable accurate and robust level-set based simulation of anisotropic wet etching and non-planar epitaxy for semiconductor fabrication. These fabrication techniques are characterized by highly crystal orientation-dependent etch/growth rates, which lead to non-convex Hamiltonians in their description by the level-set equation. As a consequence, instable surface propagation may emerge, leading to unphysical results. We propose a calibration-free Stencil Lax-Friedrichs scheme and an advanced adaptive time-stepping approach, tailored to the level-set speed functions associated with anisotropic etching and epitaxy. The scheme calculates the numerical dissipation based on information about the local geometry and the nature of the etch rates/growth function, which enables an optimized trade-off between overly rounding of sharp geometric features and stable surface propagation. Furthermore, we introduce the deposition top layer method, which allows for robust handling of multiple material regions in non-planar epitaxy simulations. Both methods are demonstrated in a prototypical implementation, which is used to validate the capability and accuracy of our approaches. In particular, two-dimensional wet etching and three-dimensional epitaxy simulations are performed and characteristic geometry parameters are compared to the ideally expected values, showing robustness and high accuracy.

MATHER measures for space–time periodic nonconvex Hamiltonians

In Gomes (Nonlinearity 15(3):581–603, 2002) developed techniques and tools with the purpose of extending the Aubry–Mather theory in a stochastic setting, namely he proved the existence of stochastic Mather measures and their properties. These results were also extended in the time-dependent setting in the doctoral thesis of the Guerra-Velasco (http://132.248.9.195/ptd2015/abril/506015252/Index.html, 2015). However, to construct analogs to the Aubry–Mather measures for nonconvex Hamiltonians, it is necessary to use the adjoint method introduced by Evans (Arch Ration Mech Anal 197:1053–1088, 2010) and Tran (Calc Var Partial Differ Equ 41:301–319, 2011); the construction of the measures is in Cagnetti et al. (SIAM J Math Anal 43(6):2601–262, citelink2011CGT). The main goal of this paper is to construct Mather measures for space–time periodical nonconvex Hamiltonians using the techniques in [10, 21] and [7]. Moreover, we also will prove that there is only one value, such that the viscous Hamilton–Jacobi equation has a smooth periodic solution unique up to an additive constant.

Stochastic homogenization of certain nonconvex Hamilton–Jacobi equations

In this paper, we prove the stochastic homogenization of certain nonconvex Hamilton–Jacobi equations. The nonconvex Hamiltonians, which are generally uneven and inseparable, are generated by a sequence of quasiconvex Hamiltonians and a sequence of quasiconcave Hamiltonians through the min-max formula. We provide a monotonicity assumption on the contact values between those stably paired Hamiltonians so as to guarantee the stochastic homogenization.

Existence and uniqueness of solutions to parabolic equations with superlinear Hamiltonians

We give a proof of existence and uniqueness of viscosity solutions to parabolic quasilinear equations for a fairly general class of nonconvex Hamiltonians with superlinear growth in the gradient variable. The approach is mainly based on classical techniques for uniformly parabolic quasilinear equations and on the Lipschitz estimates provided in [S. N. Armstrong and H. V. Tran, Viscosity solutions of general viscous Hamilton–Jacobi equations, Math. Ann. 361 (2015) 647–687], as well as on viscosity solution arguments.

A kernel based high order “explicit” unconditionally stable scheme for time dependent Hamilton–Jacobi equations

In this paper, a class of high order numerical schemes is proposed for solving Hamilton–Jacobi (H–J) equations. This work is regarded as an extension of our previous work for nonlinear degenerate parabolic equations, see Christlieb et al. [14], which relies on a special kernel-based formulation of the solutions and successive convolution. When applied to the H–J equations, the newly proposed scheme attains genuinely high order accuracy in both space and time, and more importantly, it is unconditionally stable, hence allowing for much larger time step evolution compared with other explicit schemes and saving computational cost. A high order weighted essentially non-oscillatory methodology and a novel nonlinear filter are further incorporated to capture the correct viscosity solution. Furthermore, by coupling the recently proposed inverse Lax–Wendroff boundary treatment technique, this method is very flexible in handing complex geometry as well as general boundary conditions. We perform numerical experiments on a collection of numerical examples, including H–J equations with linear, nonlinear, convex or non-convex Hamiltonians. The efficacy and efficiency of the proposed scheme in approximating the viscosity solution of general H–J equations is verified.

Hydrodynamic Limit and Viscosity Solutions for a Two‐Dimensional Growth Process in the Anisotropic KPZ Class

AbstractWe study a two‐dimensional stochastic interface growth model that is believed to belong to the so‐called anisotropic KPZ (AKPZ) universality class [4,5]. It can be seen either as a two‐dimensional interacting particle process with drift that generalizes the one‐dimensional Hammersley process [1,24], or as an irreversible dynamics of lozenge tilings of the plane [5,29]. Our main result is a hydrodynamic limit: the interface height profile converges, after a hyperbolic scaling of space and time, to the solution of a nonlinear first‐order PDE of Hamilton‐Jacobi type with nonconvex Hamiltonian (nonconvexity of the Hamiltonian is a distinguishing feature of the AKPZ class). We prove the result in two situations: (1) for smooth initial profiles and times smaller than the time Tshock when singularities (shocks) appear or (2) for all times, including t > Tshock, if the initial profile is convex. In the latter case, the height profile converges to the viscosity solution of the PDE. As an important ingredient, we introduce a Harris‐type graphical construction for the process. © 2018 Wiley Periodicals, Inc.

An Alternative Formulation of Discontinous Galerkin Schemes for Solving Hamilton–Jacobi Equations

The aim of this paper is to develop an alternative formulation of discontinuous Galerkin (DG) schemes for approximating the viscosity solutions to nonlinear Hamilton–Jacobi (HJ) equations. The main difficulty in designing DG schemes lies in the inherent non-divergence form of HJ equations. One effective approach is to explore the elegant relationship between HJ equations and hyperbolic conservation laws: the standard DG scheme is applied to solve a conservation law system satisfied by the derivatives of the solution of the HJ equation. In this paper, we consider an alternative approach to directly solving the HJ equations, motivated by a class of successful direct DG schemes by Cheng and Shu (J Comput Phys 223(1):398–415, 2007), Cheng and Wang (J Comput Phys 268:134–153, 2014). The proposed scheme is derived based on the idea from the central-upwind scheme by Kurganov et al. (SIAM J Sci Comput 23(3):707–740, 2001). In particular, we make use of precise information of the local speeds of propagation at discontinuous element interface with the goal of adding adequate numerical viscosity and hence naturally capturing the viscosity solutions. A collection of numerical experiments is presented to demonstrate the performance of the method for solving general HJ equations with linear, nonlinear, smooth, non-smooth, convex, or non-convex Hamiltonians.

The selection problem for discounted Hamilton–Jacobi equations: some non-convex cases

Here, we study the selection problem for the vanishing discount approximation of non-convex, first-order Hamilton–Jacobi equations. While the selection problem is well understood for convex Hamiltonians, the selection problem for non-convex Hamiltonians has thus far not been studied. We begin our study by examining a generalized discounted Hamilton–Jacobi equation. Next, using an exponential transformation, we apply our methods to strictly quasi-convex and to some non-convex Hamilton–Jacobi equations. Finally, we examine a non-convex Hamiltonian with flat parts to which our results do not directly apply. In this case, we establish the convergence by a direct approach.

Error analysis of the high order scheme for homogenization of Hamilton–Jacobi equation

In this paper, employing ideas developed for conservation law equations such as the Lax–Friedrich-type and Godunov-type numerical fluxes, we describe the numerical schemes for approximating the solution of the limit problem arising in the homogenization of Hamilton–Jacobi equations. All approximation methods involve three steps. The first scheme is a provably monotonic discretization of the cell problem for approximating the effective Hamiltonian for a given vector P∈RN. Next, using interpolation, we present an approximation of the effective Hamiltonian in the domain RN. Finally, the numerical schemes of the Hamilton–Jacobi equations with the effective Hamiltonian approximation are constructed. We also present global error estimates including all the discrete mesh sizes. The theoretical results are illustrated through numerical examples, including two convex Hamiltonians and two non-convex Hamiltonians.

Min–max formulas and other properties of certain classes of nonconvex effective Hamiltonians

This paper is the first attempt to systematically study properties of the effective Hamiltonian $$\overline{H}$$ arising in the periodic homogenization of some coercive but nonconvex Hamilton–Jacobi equations. Firstly, we introduce a new and robust decomposition method to obtain min–max formulas for a class of nonconvex $$\overline{H}$$ . Secondly, we analytically and numerically investigate other related interesting phenomena, such as “quasi-convexification” and breakdown of symmetry, of $$\overline{H}$$ from other typical nonconvex Hamiltonians. Finally, in the appendix, we show that our new method and those a priori formulas from the periodic setting can be used to obtain stochastic homogenization for the same class of nonconvex Hamilton–Jacobi equations. Some conjectures and problems are also proposed.

On the existence of correctors for the stochastic homogenization of viscous Hamilton–Jacobi equations

We prove, under some assumptions, the existence of correctors for the stochastic homogenization of “viscous” possibly degenerate Hamilton–Jacobi equations in stationary ergodic media. The general claim is that, assuming knowledge of homogenization in probability, correctors exist for all extreme points of the convex hull of the sublevel sets of the effective Hamiltonian. Even when homogenization is not a priori known, the arguments imply the existence of correctors and, hence, homogenization in some new settings. These include positively homogeneous Hamiltonians and, hence, geometric-type equations including motion by mean curvature, in radially symmetric environments and for all directions. Correctors also exist and, hence, homogenization holds for many directions for nonconvex Hamiltonians and general stationary ergodic media.

Homogenization of viscous and non-viscous HJ equations: a remark and an application

It was pointed out by P.-L. Lions, G. Papanicolaou, and S.R.S. Varadhan in their seminal paper (1987) that, for first order Hamilton–Jacobi (HJ) equations, homogenization starting with affine initial data implies homogenization for general uniformly continuous initial data. The argument makes use of some properties of the HJ semi-group, in particular, the finite speed of propagation. This property is lost for viscous HJ equations. In this paper we prove the above mentioned implication in both viscous and non-viscous cases. Our proof relies on a variant of Evans’s perturbed test function method. As an application, we show homogenization in the stationary ergodic setting for viscous and non-viscous HJ equations in one space dimension with non-convex Hamiltonians of specific form. The results are new in the viscous case.

Effective nonlinear Neumann boundary conditions for 1D nonconvex Hamilton–Jacobi equations

We study Hamilton–Jacobi equations in [0,+∞) of evolution type with nonlinear Neumann boundary conditions in the case where the Hamiltonian is not necessarily convex with respect to the gradient variable. In this paper, we give two main results. First, we prove for a nonconvex and coercive Hamiltonian that general boundary conditions in a relaxed sense are equivalent to effective ones in a strong sense. Here, we exhibit the effective boundary conditions while for a quasi-convex Hamiltonian, we already know them (Imbert and Monneau, 2016). Second, we give a comparison principle for a nonconvex and nonnecessarily coercive Hamiltonian where the boundary condition can have constant parts.

Algorithm for Overcoming the Curse of Dimensionality For Time-Dependent Non-convex Hamilton–Jacobi Equations Arising From Optimal Control and Differential Games Problems

In this paper we develop a parallel method for solving possibly non-convex time-dependent Hamilton---Jacobi equations arising from optimal control and differential game problems. The subproblems are independent so they can be implemented in an embarrassingly parallel fashion, which usually has an ideal parallel speedup. The algorithm is proposed to overcome the curse of dimensionality (Bellman in Adaptive control processes: a guided tour. Princeton University Press, Princeton, 1961; Dynamic programming. Princeton University Press, Princeton, 1957) when solving HJ PDE . We extend previous work Chow et al. (Algorithm for overcoming the curse of dimensionality for certain non-convex Hamilton---Jacobi equations, Projections and differential games, UCLA CAM report, pp 16---27, 2016) and Darbon and Osher (Algorithms for overcoming the curse of dimensionality for certain Hamilton---Jacobi equations arising in control theory and elsewhere, UCLA CAM report, pp 15---50, 2015) and apply a generalized Hopf formula to solve HJ PDE involving time-dependent and perhaps non-convex Hamiltonians. We suggest a coordinate descent method for the minimization procedure in the Hopf formula. This method is preferable when even the evaluation of the function value itself requires some computational effort, and also when we handle higher dimensional optimization problem. For an integral with respect to time inside the generalized Hopf formula, we suggest using a numerical quadrature rule. Together with our suggestion to perform numerical differentiation to minimize the number of calculation procedures in each iteration step, we are bound to have numerical errors in our computations. These errors can be effectively controlled by choosing an appropriate mesh-size in time and the method does not use a mesh in space. The use of multiple initial guesses is suggested to overcome possibly multiple local extrema in the case when non-convex Hamiltonians are considered. Our method is expected to have wide application in control theory and differential game problems, and elsewhere.