14
Scheduling Tian He Frank Zhi TV Commercial

Scheduling

  • Upload
    raven

  • View
    44

  • Download
    0

Embed Size (px)

DESCRIPTION

Scheduling. TV Commercial. Tian He Frank Zhi. TV Commercial Scheduling Background. Commercials traditionally scheduled by hand Human error Case study: NBC network Automated scheduling heuristics Increased profit. Scheduling Heuristic considerations. - PowerPoint PPT Presentation

Citation preview

Page 1: Scheduling

SchedulingTian HeFrank Zhi

TV Commercial

Page 2: Scheduling

TV Commercial SchedulingBackground

• Commercials traditionally scheduled by hand– Human error

• Case study: NBC network– Automated scheduling

heuristics– Increased profit

Page 3: Scheduling

• Number of commercials bought by a company

• First and last slots in a commercial break highly coveted

• Evenly spaced commercials of same type

• Juxtaposition of competitor commercials in same break

SchedulingHeuristic considerations

Page 4: Scheduling

• General MIP Model– Considers commercial types, number bought, commercial and break lengths, client goals– Enormous runtime!

• ISCI model (integer program)– Evenly spaced commercials of same type– Intuitive but has problems

• Network Flow model– Improves runtime greatly for smaller problems– Less intuitive

Heuristics

Page 5: Scheduling

General MIP- variables

Takes into account client goals

Page 6: Scheduling

General MIP –model

• Client goals– Product conflict constraints– Position percentage

constraints

Page 7: Scheduling

ISCI model - variables

Page 8: Scheduling

ISCI model: IP

The method is simple and intuitive, but for larger problems becomes overwhelming.

Page 9: Scheduling

ISCI model: Colored Balls

• We coded the ISCI model.– example uses I=3, with i1=3, i2=2, i3=4

• IP solves schedule optimally as follows:

• I eyeball schedule as follows:

Page 10: Scheduling

Conclusion of ISCI problem

• VERY small spacing problems can be done by hand

• Larger scale problems are infeasible by hand and require heuristics like the ISCI model

Page 11: Scheduling

Network FlowFormulation - variables

Page 12: Scheduling

Network FlowFormulation

• Each commercial type has above networkNodes: slot spacesArcs: f(c,j,k) indicates that a commercial of type “c” is consecutively scheduled in

slots “j” and “k”

Page 13: Scheduling

Conclusion

• Scheduling heuristics much more efficient than scheduling by hand “on-the-fly”

• Important to find heuristics that will solve large problems in an acceptable timeframe

Page 14: Scheduling

References• Bussieck, Michael R. Bollapragada, Srinivas. Mallik,

Suman. "Scheduling Commercial Videotapes in Broadcast Television." Operations Research 52 (2004): 679-89

• Garbiras, Marc. Bollapragada, Srinivas. "Scheduling commercials on Broadcast Television." Operations Research 52 (2004): 337-45.