Competitive Networks and Measures of Intransitivity
Competitive Networks and Measures of Intransitivity
复制标题
竞争网络和不及物性测量
DOI:
10.1086/283539
复制
发表时间:
1979
期刊:
影响因子:
--
通讯作者:
P. Petraitis
中科院分区:
文献类型:
--
作者:
P. Petraitis
While conjectures about the importance of intransitive competitive relationships are not new (Lewontin 1968), recent work by several people (Gilpin 1975; Jackson and Buss 1975; Jackson 1977; May and Leonard 1975) has brought he problem to the attention of ecologists. Yet, intransitivity in competitive communities cannot be examined without some measure of the relative level of intransitivity among competitors. Two measures, one proposed by Kendall and Smith (1940) and the other by Landau (1951), have been used for this purpose. These measures, however, are not appropriate for this problem. Kendall and Smith confine their attention to intransitive triads (A beats B, B beats C, but C beats A). Landau's measure is the same as Kendall and Smith's measure for an odd number of objects. However, the number of intransitive relationships is not a simple function of the number of intransitive triads. An example is given in table 1. In the first wo cases there are eight intransitive triads. Yet case 1 requires four reversals, while case 2 requires three reversals, to be transformed into a hierarchy.A more desirable measure of intransitivity in competitive systems would be based on the minimum number of reversals required to transform a network into a hierarchy and would be standardized by the maximum number of intransitive relationships possible. For such a measure, let t= 1-sIM where s is the minimum number of required reversals and M is the maximum number of possible intransitive relationships. The range of t is from zero to one; when t= 1 there are no intransitive relationships.