|
|
|
|
|
|
|
A brief look at the enormous times taken by enumerative plans to discover fit structures, even when the domain of action  is greatly constrained, makes it clear that efficiency is a sine qua non of studies of adaptation. Efficiency acts as a cutting edge, shearing away plans "too slow" to serve as hypotheses about natural systems or as algorithms for artificial systems. Whether an adaptive plan is to serve as hypothesis or algorithm, information about its robustnessits efficiency in the environments eis critical. The latter part of this book will be much concerned with this topic. Chapter 4 introduces a critical tool for the investigation and construction of efficient adaptive plansschemata. This generalization of coadapted sets of alleles provides an efficient way of defining and exploiting properties associated with above-average performance. Chapter 5 develops a criterion for measuring the efficiency with which adaptive plans improve average performance and then relates this criterion to the exploitation of schemata. Chapter 6 introduces generalized genetic plans and chapter 7 establishes their robustness. Chapter 8 studies mechanisms which enable genetic plans to use predictive modeling for flexible exploitation of the large fluxes of information provided by typical environments. |
|
|
|
|
|
|
|
|
The emphasis throughout the book is on general principles which help to resolve the problems and questions raised in this chapter. One particular interest will be the solution of problems involving hundreds to hundreds of thousands of interdependent parameters and multitudes of local optimaproblems which largely lie outside the prescriptions of present day computational mathematics. |
|
|
|
|
|