• No results found

Appendix 3 Overview of interviews ... 10 Appendix 4 Overview heuristic ... 11 Appendix 5 Timetables for examples penalty calculation ... 21 Appendix 6 Experiments ... 26

5

APPENDIX 1

GLOSSARY

Symbol Term Description

a solution Solution of a scheduling problem, consisting of an assignment to a time and place for each event a(e) assignment Assignment of event e to a room and timeslot

ã candidate solution Solution that may replace the current solution. a* solution optimal The best solution to the scheduling problem

β cooling factor Constant that determines how much the temperature is changed each time c class Index used for classes, the set of all possible timeslots is C. c(e) is the sets of classes c that attend event e cacc_all Counter for the total number of solutions that is accepted (all operators combined).

cacc_o Counter for the number of times a solution generated with operator o is accepted. cnew_o Counter for the number of times operator o generates a new and feasible solution.

ctot_o Counter for the total number of times operator o is selected. δ solution value difference in Difference between the solutions values of the current solution a and candidate solution ã δ(x,y) Binary variable δ(x,y)=1 if x=y, δ(x,y)=0 if x≠y

e event Educational activity, e.g. a lecture or tutorial. The set of all events in the problem is E experiment Evaluation of a scenario. Consists of 40 to 100 runs under the same constraints with a varying requested utilisation or

frequency.

F frequency Fraction of rooms in a building that is used

FA frequency achieved Fraction of rooms in a building that is used for events that are included in a timetable after solving the scheduling FC frequency critical Frequency above which it is almost impossible to find a feasible timetable FR requested frequency Fraction of rooms in a building that is used if all events are included in a timetable iterationc iteration current Indicator of the current iteration of the SA heuristic

iterationmax maximum iteration The number of iterations after which the SA heuristic stops l lecturer Index used for lecturers, the set of all possible timeslots is L. l(e) is the sets of lecturers l that attend event e lp learning period Number of iterations after which the performance of local operators is evaluated and the heuristic learns which

operators to use in the coming period.

6

Symbol Term Description

ni occurence penalty Constant, the number of times constraint i is violated by a solution. o local operator A low level heuristic that creates candidate solutions by altering the current solution. O Occupancy Fraction of seats-hours in a buidling that is used, limited to roomslots that are used. P penalty matrix Matrix that indicates if a full penalty is awarded for each given room and event type mismatch. pi penalty Constant, the number of penalty points awarded per violation of constraint i. pro operator selection

probability

Probability that local operator o will be used to create candidate solution ã.

π1 (a(e)) room First variable of the assignment of event e, its location. π2 (a(e)) timeslot Second variable of the assignment of event e, its time.

r room Index used for rooms, the set of all possible rooms is R. rmin acceptance minimum

ratio

Constant that determines when a reheating period is started.

roomslot Measure of capacity of educational space, multiplication of rooms and time. run One instance of solving a scheduling problem by running the SA heuristic. S solution space Set of all possible solutions to a scheduling problem.

seat-hour Measure of capacity of educational space, multiplication of seats and time. t timeslot Index used for timeslots, the set of all possible timeslots is T. tcurrent temperature current The temperature that is used to determine if a candidate solution is accepted.

tend temperature end Constant that will be the final t_current if no reheating phases occur during a run of the heuristic. timpr improvement temperature

The last temperature at which a solution was found that was an improvement of the previous solution, the current temperature may revert to this value from time to time. tst temperature start Constant that is used as the initial t_current and as the maximum for t_current while reheating. U unavailability matrix Matrix that indicates if a staff member is available to teach during a timeslot. U utilisation Fraction of seats-hours in a buidling that is used.

7

Symbol Term Description

UA utilisation achieved Fraction of seats-hours in a buidling that is used for events that are included in a timetable after solving the scheduling problem.

UC utilisation critical Utilisation above which it is almost impossible to find a feasible timetable. UR requested utilisation Fraction of seats-hours in a buidling that is used if all events are included in a timetable. wmin weight factor minimum Constant, if the calculated wo is lower than wmin, we use wmin instead of wo.

wo weight factor Weight factor of operator o, used to calculate pro.

Acronym Description

AUAS Amsterdam University of Applied Sciences BE Business Economics, educational programme CB-CTT Curriculum based course timetabling

FA Finance & Accounting, a cluster of programmes within the School Economics & Management FMB Fraijlemaborg, the building that houses FA and IBS

HTT Highschool timetabling

IBS International Business School, a cluster of programmes within the faculty of Economics & Management ITC International Timetabling Competition

SA Simulated annealing, a type of heursistic

8

Related documents