Numerical detection of instability regions for delay models with delay-dependent parameters (Q885934): Difference between revisions

The paper studies a criterion for the occurence of Hopf bifurcations in delay differential equations (DDEs) and demonstrates its use for three DDEs arising in population dynamics. The authors advertise their criterion for numerical computations. The reviewer does not recommend (at least for higher dimensional systems) this approach because the criterion is based on the characteristic quasi-polynomial, which is known to be dangerously unstable compared to standard methods as employed for DDE-Biftool [see \textit{D. Roose, T. Luzyanina, K. Engelborghs} and \textit{W. Michiels}, ``Software for stability and bifurcation analysis of delay differential equations and applications to stabilization'', Niculescu, Silviu-Iulian (ed.) et al., Advances in time-delay systems. Berlin: Springer. Lecture Notes in Computational Science and Engineering 38, 167--181 (2004; Zbl 1065.34077)]. This standard package is also able to treat problems where other system parameters depend on the delay (by adding algebraic constraints), contrary to the authors' claim. How the standard tools can be applied to problems with distributed delays for certain common integral kernels is explained in [\textit{T. Luzyanina, D. Roose} and \textit{K. Engelborghs}, Appl. Numer. Math. 50, 75--92 (2004; Zbl 1052.65119)]. Interestingly, in all of the examples discussed in the paper the delays relevant for stability could have been scaled to \(1\) without loss of generality by rescaling time. In summary, the most interesting part of the paper is the discussion and introduction of the biological models. The reviewer fails to see the additional numerical or conceptual difficulty, as perceived by the authors, due to the dependence of system parameters on the delay.