Kot Mark, Phillips Austin
Department of Applied Mathematics, Box 353925, University of Washington, Seattle, WA 98195-3925, USA.
Quantitative Ecology & Resource Management, Box 352182, University of Washington, Seattle, WA 98195-2182, USA.
Math Biosci. 2015 Apr;262:65-72. doi: 10.1016/j.mbs.2014.12.007. Epub 2015 Jan 30.
Integrodifference equations have recently been used as models for populations undergoing climate-driven habitat movement. In these models, the persistence of a population is governed by the maximal or dominant eigenvalue of a Fredholm integral equation with an asymmetric kernel; this eigenvalue determines the critical translational speed for extinction of the population. Since direct methods for finding eigenvalues are often analytically or computationally expensive, we explored the extensive literature on alternative methods for localizing maximal eigenvalues. We found that a sequence of iterated row sums provide upper and lower bounds for the maximal eigenvalue. Alternatively, arithmetic and geometric symmetrization yield upper and lower bounds. Geometric symmetrization is especially valuable and leads to a simple Rayleigh quotient that can be used to analytically approximate the critical-speed curve. Our research sheds new light on the interpretation and limitations of the average-dispersal-success approximation; it also provides a generalization of this useful tool for asymmetric kernels.
积分差分方程最近被用作经历气候驱动栖息地迁移的种群模型。在这些模型中,种群的持续性由具有非对称核的弗雷德霍姆积分方程的最大或主导特征值决定;该特征值确定了种群灭绝的临界平移速度。由于直接求特征值的方法通常在分析上或计算上成本很高,我们研究了关于定位最大特征值的替代方法的大量文献。我们发现,一系列迭代行和为最大特征值提供了上下界。另外,算术和几何对称化也能得到上下界。几何对称化特别有价值,并导致一个简单的瑞利商,可用于解析近似临界速度曲线。我们的研究为平均扩散成功近似的解释和局限性提供了新的见解;它还对这个适用于非对称核的有用工具进行了推广。