Numerical Algorithms ( IF 1.7 ) Pub Date : 2020-08-20 , DOI: 10.1007/s11075-020-00992-9 Igor Omelyan , Yuri Kozitsky , Krzysztof Pilorz
An algorithm is proposed for finding numerical solutions of a kinetic equation that describes an infinite system of point particles placed in \(\mathbb {R}^{d} (d \geq 1)\). The particles perform random jumps with pair-wise repulsion in the course of which they can also merge. The kinetic equation is an essentially nonlinear and nonlocal integro-differential equation, which can hardly be solved analytically. The numerical algorithm which we use to solve it is based on a space-time discretization, boundary conditions, composite Simpson and trapezoidal rules, Runge-Kutta methods, and adjustable system-size schemes. We show that, for special choices of the model parameters, the solutions manifest unusual time behavior. A numerical error analysis of the obtained results is also carried out.
中文翻译:
具有合并和排斥跳跃的空间种群动力学模型动力学方程数值解的算法
提出了一种寻找动力学方程数值解的算法,该算法描述了置于\(\ mathbb {R} ^ {d}(d \ geq 1)\)中的点粒子的无限系统。粒子以成对排斥的方式执行随机跳跃,在此过程中它们也可以合并。动力学方程是一个本质上是非线性且非局部的积分微分方程,很难通过解析来求解。我们用来解决它的数值算法是基于时空离散,边界条件,辛普森和梯形规则,Runge-Kutta方法和可调系统大小方案。我们表明,对于模型参数的特殊选择,解决方案表现出异常的时间行为。还对获得的结果进行了数值误差分析。