SGA search dynamics on second order functions [chapter]

Bart Naudts, Alain Verschoren
1998 Lecture Notes in Computer Science  
By comparing its search dynamics to that of a simple O(n 2 ) heuristic, we are able to analyze the behavior of the simple genetic algorithm on second order functions, whose optimization is shown to be an NP-equivalent problem. It appears that both algorithms approach these tness functions in the same, intuitive way: they start by deciding the obvious and most probable, and then they proceed to make more di cult decisions. Useful information about the optimization problem is, among others,
more » ... ed by statistical physics: lattice gases can be modeled as second order functions.
doi:10.1007/bfb0026602 fatcat:77xex6hhsrbofislwox6dgaz5i