ABSTRACT
Because complexity is seen to exist in many diverse, natural, and man-made systems, it has been very difŸcult to agree on common, working deŸnitions for complex systems (CSs) and, consequently, to agree on a universal set of quantitative descriptors for their complexity. In fact, universal descriptors may not exist, and probably should not exist. Complexity measures have thus evolved ad hoc for various classes of CSs (i.e., genomic, biochemical, physiological, ecological, economic, etc., as well as computer programs, data structures, and time series). Note that any quantitative measure of complexity must be graded, thus an arbitrary threshold criterion must be used to compare the complexity “score” of CS “A” with CS “B.” Some CSs are called very complex by acclaim (e.g., the adaptive, complex human CNS), while the levels of complexity of others can be argued (e.g., the human complement system vs. the human cardiovascular system). It is the graphical models of CNLSs to which we generally apply complexity measures.
