I found some other explanation on xenon oscillation in a paper "The Influence of Xenon-135 on Reactor Operation" written by Paul L. Roggenkamp
"The oscillations can be damped, in which case each succeeding cycle diminishes, or undamped, in which case the amplitude of each succeeding cycle increases. The dividing line or threshold in terms of power density) for undamped oscillations depends on several factors that are characteristics of a particular reactor charge.Characteristics that tend to lower the threshold are: lower fuel loading (e. g., end of fuel cycle vs.beginning); greater flatness of the power distribution(e. g., greater number of fuel assemblies operating at the same power); and smaller negative temperature coefficients of reactivity.Because of the large size, high power density, and great degree of power distribution flatness,most reactor charges that have been operated at SRS are susceptible to xenon oscillations despite their negative temperature coefficients of reactivity."
One question trouble me also: "why the wave with loosely coupled peaks is more readily to oscillate than tight coupled peaks " or "why the wave with two peak at the both end is more readily to oscillate than cosine wave"? Is there any hankbook or website analysing these issues?