next up previous
Next: References Up: No Title Previous: Conclusions

Acknowledgments

We would like to thank Andrew Gee and Martin Simmen for helpful comments and suggestions, Seong-Joon Yoo for useful discussions on the preprocessing phase, and Hon Yau for comments and suggestions on the paper. The first author is very grateful for the valuable discussion and help of Robert Irwin in converting and formatting the registration data prior to the scheduling process. Many thanks go to Karen Bedard for providing us with the data and answering so many questions we had about it, Meg Cortese for providing us with a set of building constraints for various departments, and Ben Ware, Vice President for Research and Computing at Syracuse University, for his support and encouragement.

  
Table 1: Size of the data set for each of the three semesters.

  
Table 2: Sparseness ratios for the data sets for each of the three semesters. Lower values indicate a harder problem.

  
Table 3: Percentage of classes scheduled and student preferences satisfied. The averages and highest and lowest values for simulated annealing were obtained using 10 independent runs. No preprocessor was used for simulated annealing.

  
Table 4: Percentage of classes scheduled and student preferences satisfied. The averages and highest and lowest values were obtained using 10 independent runs. An expert system was used as a preprocessor for the simulated annealing.

  
Figure 3: Cost vs. log (base 2) of temperature for simulated annealing using geometric and adaptive cooling.

  
Figure 4: Cost vs. number of annealing steps for adaptive cooling with reheating and geometric cooling.

   figure623
Figure 5: Average specific heat vs. log (base 2) of temperature.

  
Figure 6: Average rate of acceptance vs. log (base 2) of temperature for annealing using adaptive cooling.

   figure637
Figure 7: Initial annealing temperature vs. average final cost for adaptive cooling with reheating and geometric cooling.

  
Figure 8: Final cost vs. the value of the parameter a, which governs the rate of cooling in the adaptive cooling schedule.

  
Figure 9: Final cost vs. the value of K, which governs the amount of reheating when using reheating as a function of cost.


next up previous
Next: References Up: No Title Previous: Conclusions

Saleh Elmohamed
Thu Sep 4 11:43:55 EDT 1997