**4. Parameter tuning**

Many researchers such De Jong, Grefenstte, Schaffer and others have contributed considerable efforts into finding the parameters values which are good for a number of numerical test problems. The evolution of the bidding strategies by Anthony and Jennings (Anthony & Jennings, 2002) employed a fixed crossover and mutation probability based on the literatures. However, these recommended values may not perform at its best in the genetic algorithm as it has been proven that the parameter values are dependent on the nature of problems to be solved (Engelbrecht, 2002). In this experiment, the crossover and mutation rates are fine tuned with different combination of probabilities in order to discover the best combination of genetic operators' probabilities. Thus, the main objective of this experiment is to improve the effectiveness of the bidding strategies by "hand tuning" the values of the crossover rate and mutation rate to allow a new combination of static crossover and mutation rates to be discovered. By improving the algorithm, more effective bidding strategies can be found during the exploration of the solution.

The experiment is subdivided to two parts. The first one varies the crossover rate and the second one varies the mutation rate. At the end of this experiment, the combination rate discovered is compared and empirically evaluated with the bidding strategies evolved in Anthony's work (Anthony, 2003).
