Lax–Hopf Based Incorporation of Internal Boundary Conditions Into Hamilton-Jacobi Equation. Part II: Computational Methods

Christian Claudel, Alexandre M. Bayen

IEEE Transactions on Automatic Control · 2010 · 139 citations · 39 references

Concepts

TL;DR

The article introduces a new method for explicitly computing solutions to a Hamilton–Jacobi PDE with piecewise affine initial, boundary, and internal conditions. Using viability theory and a Lax–Hopf formula, the authors assemble analytical contributions of each affine condition into a Lax–Hopf algorithm that evaluates the solution at any time via a semi‑analytical expression, and benchmark its runtime against a Godunov scheme. The semi‑analytical method achieves machine‑level accuracy, matching the precision of finite‑difference schemes while avoiding their discretization error.

Abstract

This article presents a new method for explicitly computing solutions to a Hamilton-Jacobi partial differential equation for which initial, boundary and internal conditions are prescribed as piecewise affine functions. Based on viability theory, a Lax-Hopf formula is used to construct analytical solutions for the individual contribution of each affine condition to the solution of the problem. The results are assembled into a Lax-Hopf algorithm which can be used to compute the solution to the partial differential equation at any arbitrary time at no other cost than evaluating a semi-analytical expression numerically. The method being semi-analytical, it performs at machine accuracy (compared to the discretization error inherent to finite difference schemes). The performance of the method is assessed with benchmark analytical examples. The running time of the algorithm is compared with the running time of a Godunov scheme.

References

39