Real Options Approach to Finding Optimal Stopping Time in Compact Genetic Algorithm

Abstract

The real options technique has emerged as an evaluation tool for investment under uncertainty. It explicitly recognizes future decisions, and the exercise strategy is based on the optimal decisions in future periods. The real options approach has been applied to many economic and financial problems, but few are in computer science and engineering. The novelty of this work lies in applying real options to a computational problem. This paper proposes using the real options technique to find an optimal stopping decision for the compact genetic algorithm. The compact genetic algorithm, a kind of genetic algorithms, represents the population as a probability distribution over a set of solutions. This distribution automatically captures the underlying uncertainty of the problem, which can be simulated to obtain an evolutionary process of the algorithm. The experiments show preliminary results of employing the real options approach to determine the optimal stopping time for the compact genetic algorithm. The proposed technique can be applied to analyze other machine-learning algorithms, such as neural networks or other variations of genetic algorithms.