Resources (for a lecture):

• Room
• Teacher
• Class
• Student
• Time
• Subject

Room:

• ID
• Name/description
• Room types (generic, laboratory, kitchen, office)
• Room size/number of places
• Room occupancy (["xx:xx - xx:xx"])

Time:

• Additional time between the hours (switching room, break, ventilation, discussions)
• Teachers/students need breaks when it is convenient (or at a fixed time)

Teacher:

• Name
• Type (teacher, professor, external employee)
• Subjects
• Availability

Students:

• Name
• classes/courses

Action/query:

• Is the resource available? (Teacher, room, class) for a period
• Set/give priority
• Set hard/soft constraint

Example rules:

• Teacher + class + student can only be in one room at the same time
• Subject can only take place in a certain room
• Subjects should be in a certain order during the week
• Minimize time between the hours
• Reduce walking time between different rooms/facilities

Algorithm

A metaheuristic evolutionary algorithm for creating timetables

Variables (G = Generations, P = Population, T = Stop Threshold, S = Selection percentage, M = Mutation probability)

• Loads all resources
• Loads all rules and sorts them by priority
• Repeat until no improvement in performance is found after T runs or the user manually stops
  • Repeat G times:
    • Repeat P times:
      • Starts with a random table or takes the one of the previous run
      • Assigns performance 0 to this run
      • Execute every rule for every resource
        • If the rule is fulfilled, its priority is added to the performance
        • If the rule is violated, its priority is subtracted from the service
        • If the rule is violated and it is absolutely necessary, the service is set to 0 and this cycle is ended
  • Saves and sorts the results according to the performance of the runs
  • Selection (S percent of the best runs are selected)
  • Recombination (some entries are swapped for two of the selected runs)
  • Mutation (some entries are swapped with a probability of M)
  • Population (100% - S are also filled in randomly generated tables)

Schema:

[
	{
		"and": [
			{
				"or": [
					{
						"equals": "1==1"
					},
					{
						"equals": "1==0"
					}
				]
			},
			{
				"equals": "1==1"
			}
		]
	}
]

{
	"ressources": {
		"lehrer": {}
	}
}