Green Banana Optimal Task Scheduler
Green Banana (GB for short) optimal task scheduler aims at finding optimal schedules for the task scheduling problem on homogeneous
multi-processor systems with communication delays.
GB is designed for users to choose from a suite of Mixed Integer Linear Programming (MILP) formulations to solve the scheduling
problem with CPLEX Optimisation Solver.
The task's execution times, the precedence constraints between tasks and communication delays are conveniently modelled with the Graph eXchange
Language (GXL) that is designed to be a standard exchange format for graphs. More information may be found on the
GXL Home Page. Example of an input GXL file can be found
here
and the corresponding output GXL file with the full schedule
here. If the input GXL file starts with a node
number of 0, the output GXL files have 1 added to every node number in-order to maintian numbering consistency.
The expected input and output GXL file formats used are found
here and
here.
The GB scheduler suite supports the following formulation and features:
- Feature batchprocess: This feature can be used to create a script to batch process all the .lp extension files with CPLEX. This essentially
appends to batch script file the sequence of steps seen in Example 1 and Example 2, given the lp filename and the timelimit. When the GB scheduler is
downloaded, type java gb -help to see an example on how this works.
- Feature help: After downloading the GB scheduler, type java gb -help to see what the Green Banana suite supports, instructions and examples on how
to use it.
- Feature gxlout: This feature allows the user to convert the CPLEX log file to GXL output format, given the CPLEX log file and the input
GXL file. To see an example, type java gb -help
- Feature printschedule: This feature prints on screen the complete schedule, given the CPLEX log file and the input
GXL file. To see an example, type java gb - help
Download
Download the Green Banana optimal task scheduler from
here.
Graphs and Results
Graphs and optimal schedules are found
here.
Reference
[1]. Sarad Venugopalan and Oliver Sinnen.
Optimal linear programming solutions for multiprocessor
scheduling with communication delays.12th International Conference, ICA3PP 2012. In Yang Xiang, Ivan Stojmenovic, BernadyO. Apduhan,
Guojun Wang, Koji Nakano, and Albert Zomaya, editors, Algorithms and Architectures for Parallel Processing, volume 7439 of Lecture Notes
in Computer Science, pages 129–138. Springer Berlin Heidelberg, 2012.
[Springer Abstract] .
[3]. Sarad Venugopalan and Oliver Sinnen.
ILP formulations for Optimal Task Scheduling with Communication Delays on Parallel Systems,
IEEE Transactions on Parallel and Distributed Systems. Accepted 1st February 2014.DOI: 10.1109/TPDS.2014.2308175
[IEEE TPDS Early Access] .
[3]. T. Davidovic, L. Liberti, N. Maculan, and N. Mladenovic. Towards the Optimal solution of the Multiprocessor Scheduling Problem
with Communication Delays. In 3rd Multidisciplinary International Conference on Scheduling: Theory and Application., pages 128–135, 2007.
[CiteSeerX Abstract] .