Models and algorithms have been implemented in GAMS (distribution 20.8 and higher). The collection of GAMS models implements P2, P3, Delta P3, Batch P3 and the Assignment Heuristic. The various options for the models and the algorithms can be passed on through the command line. For example,
gams cc --curid=12 --alg=ah
All options are also described in the header of the GAMS model cc.gms:
This model computes an optimal or near optimal schedule for the ISCI problem: Major advertisers purchase several slots to air commercials during a given time period on a broadcast network. We study the problem of scheduling advertiser's commercials in the slots it purchased when the same commercial is to be aired multiple times. Under such a situation, the advertisers typically want the airings of a commercial to be as much evenly spaced as possible. Following options are possible: curid - Identifier of the problem instance (0-40) alg - Algorithm/Model used to solve instance: p2 GAMS/CPLEX P2 p3 GAMS/CPLEX P3 d3 P3/Delta Heuristic b3 P3/Batching Heuristic ah Assignment Heuristic resmax - Maximum time available to solve the instance preplaced - use instance with preplaced balls (not available with alg=ah)By default the model runs the Batch Heuristic (b3) for instance 1 with a maximum of 600 CPU second.
The model creates appends to a log file
ID: 34 p2 SOBJ: 298.4941 BND: 112.2695 T: 625sThe instance 34 has been solve with model P2. The model found a solution of value 298.4941 and established a lower bound of 112.2695 and ran for 625 seconds.
Please download the Excel Spreadsheet with result for running P2, P3, Delta P3 and the Batch P3 Heuristic for the instances with preplaced balls.
In order to run the models you need a license for GAMS and GAMS/CPLEX.