Abstract:This paper studies a family of convolution quadratures, a numerical technique for efficient evaluation of convolution integrals. We employ the block generalized Adams method to discretize the underlying initial value problem, departing from the well-established approaches that rely on linear multistep formulas or Runge-Kutta methods. The convergence order of the proposed convolution quadrature can be dynamically controlled without requiring grid point adjustments, enhancing exibility. Through strategic selection of the local interpolation polynomial and block size, the method achieves high-order convergence for calculation of convolution integrals with hyperbolic kernels. We provide a rigorous convergence analysis for the proposed convolution quadrature and numerically validate our theoretical findings for various convolution integrals.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is to develop a new convolution quadrature method to calculate the convolution integral (CI) efficiently and accurately, especially when dealing with integrals with hyperbolic kernels. Specifically, the goal of the paper is to combine the advantages of the linear multistep convolution quadrature (LMCQ) and the Runge - Kutta convolution quadrature (RKCQ) and propose a convolution quadrature method (BGACQ) based on the block - generalized Adams method (BGA). This method aims to:
1. **Implementation on a uniform grid**: It does not need to rely on non - uniform nodes like RKCQ, which simplifies the calculation.
2. **Flexible adjustment of the convergence order**: The convergence order can be dynamically adjusted without changing the grid points, similar to LMCQ.
3. **High - order and stable calculation**: For convolution integrals with hyperbolic kernels, it can achieve high - order convergence and maintain numerical stability.
### Specific problem description
The convolution integral (CI) has wide applications in many fields, such as time - domain boundary integral equations, Volterra convolution - type integral equations, fractional - order differential equations and Maxwell equations, etc. When the kernel function \(k(s)\) is known, there are many effective methods to calculate CI, such as the convolution spline method and the spectral method based on the orthogonal polynomial convolution matrix. However, if only the Laplace transform \(K(\lambda)\) of the kernel function is known, then convolution quadrature (CQ) becomes an attractive option because it is suitable for stable integration over long time intervals.
### Main contributions of the paper
The paper develops a new convolution quadrature method (BGACQ) by introducing the block - generalized Adams method (BGA) to discretize the initial value problem (IVP). The main features include:
- **High - order convergence**: By choosing appropriate local interpolation polynomials and block sizes, BGACQ can achieve high - order convergence.
- **Flexibility**: The convergence order can be adjusted without changing the grid points.
- **Numerical stability**: Through strict convergence analysis, the A - stability of BGACQ is proved, and its theoretical results on various convolution integrals are verified.
### Mathematical formula representation
The key formulas involved in the paper are as follows:
- Definition of the convolution integral:
\[
(K(\partial_t)g)(t):=\int_0^t k(s)g(t - s)\,ds
\]
- Initial value problem (IVP):
\[
y'(t)=\lambda y(t)+g(t),\quad y(0) = 0
\]
- Discretized convolution quadrature formula:
\[
U_n=\sum_{j = 0}^n W_jG_{n - j}
\]
These formulas show how to discretize continuous problems and solve convolution integrals by numerical methods.
### Conclusion
Through strict mathematical derivations and numerical experiments, the paper proves the effectiveness of BGACQ in calculating convolution integrals. Especially for convolution integrals with hyperbolic kernels, BGACQ not only achieves high - order convergence but also maintains numerical stability. This provides new tools and methods for solving practical problems in related fields.