Year of Award


Document Type

Thesis - Campus Access Only

Degree Type

Master of Science (MS)

Degree Name

Computer Science

Department or School/College

Department of Computer Science

Committee Chair

Alden Wright

Commitee Members

Jesse Johnson, Woodam Chung


evolutionary algorithms, NK landscape, speed of evolution, varying environments


University of Montana


One of the most important tasks in computer science and artificial intelligence is optimization. Computer scientists use simulation of natural evolution to create algorithms and data structures to solve complex optimization problems. This field of study is called evolutionary computation. In evolutionary computation, the speed of evolution is defined as the number of generations needed for an initially random population to achieve a given goal. Recent studies have shown that varying environments might significantly speed up evolution, and suggested modularly varying goals can accelerate optimization algorithms. In this thesis, we study the effect of varying goals on the speed of evolution. Two test models, the NK model and the midunitation model, are used for this study. Three different evolutionary algorithms are used to test the hypothesis. Statistical analyses of the results showed that under NK model, evolution with fixed goal is faster than evolution with switching goals. Under midunitation model, different algorithms lead to different results. With some string lengths using hill climbing, switching goals sped up evolution. With other string lengths using hill climbing, and using the other evolutionary algorithms, either evolution with a fixed goal was faster or results were inconclusive.

This record is only available
to users affiliated with
the University of Montana.

Request Access



© Copyright 2009 Zhongmiao Xiao