Development of a Taxi Planning Tool Using Genetic Optimization

摘要

Within the framework of the joint DUT/FAE and NLR project concerned with A-SMGCS,211u001ea taxi-planning tool was developed. To do this, an optimization method called 211u001eGenetic Optimization was used. This method is based on Darwins evolution theory, 211u001estating that successive generations of a population are better adapted to the 211u001eenvironment in which the population exists. In the Genetic Optimization method, a 211u001egroup of solutions called a population is created, that is reproduced in a way 211u001emodeled on the natural world, and in this manner new and (hopefully) better 211u001esolutions are created. A number of algorithms based on this method is available 211u001ein the public domain, and some initial tests were done using two different 211u001eprograms, GENESIS and GENOCOP. After the initial tests, it was concluded that 211u001eGENOCOP seemed to be the most promising algorithm for this particular program, 211u001emainly based on the fact that it can handle constraints and that it uses floating 211u001epoint representation, which makes it easier to make the necessary adjustments to 211u001ethe program. Using this program, a tool was developed and a limited optimization 211u001eof the parameters of the program was performed. With the final version of the 211u001eprogram, a comparison was made with a similar tool based on LP and MIP methods. 211u001eDue to the differences in the modeling, no comparison could be made between the 211u001eperformance of the two tools, but some qualitative comparisons were made.