Evolutionary computation of a deterministic switching regressions estimator

Summary  A deterministic switching regressions estimator is evaluated using an evolutionary method based on genetic algorithms. Distinctive aspects of the method include (1) a combination of simple and random chromosomal crossover and (2) extension of the principle of natural selection to internal parameterization. The evolutionary computation duplicates, significantly faster, the results of an existing enumerative method in samples small enough to permit enumeration. It also provides the ability to calculate the estimator in much larger sample sizes than is possible with the enumerative approach. An example problem from the United States gasoline market is given. Funding support provided by the Gerondelis Foundation, Inc. is gratefully acknowledged.