离散化
数学
趋同(经济学)
指数函数
应用数学
网格生成
数值分析
有限差分
多边形网格
边界(拓扑)
收敛速度
有限差分法
杠杆(统计)
花键(机械)
数学分析
一致收敛
领域(数学分析)
边值问题
自适应网格优化
数学优化
奇异摄动
积分方程
时间离散化
自适应算法
有限元法
边界层
作者
Aditya Kaushik,Mamta Godara,Manju Sharma
摘要
Abstract This paper introduces a novel hybrid difference approximation for solving time-dependent reaction-diffusion problems with shifts and integral boundary conditions. The problem is singularly perturbed, and its solution exhibits multiscale behaviour. The approach uses a backward difference discretization in time on a uniform mesh. A key feature is the generation of an adaptive moving mesh, partitioning the spatial domain to leverage the strengths of different discretization methods. It combines a cubic spline difference method within the boundary layer region and an exponential spline difference method in the outer layer region. The mesh is generated using the equidistribution principle. This approach enhances the accuracy of numerical solutions while maintaining computational efficiency. The paper also provides a thorough theoretical analysis and numerical results for four model problems. Numerical experiments confirm parameter uniform convergence and support the theoretical outcomes.
科研通智能强力驱动
Strongly Powered by AbleSci AI