Dual-Period Task Scheduling

MATLAB/Simulink simulation of the paper 'Exploiting a Dual-Mode Strategy for Performance-Maximization and Resource-Efficient CPS Design'.

Requirements

1. Install MATLAB (c) R2017a or later
2. Install MinGW-w64 compiler (DO NOT include any space in the installation folder!)

• for MATLAB R2015b, R2016a, R2016b, R2017a: MinGW GCC 4.9.2 from TDM (https://sourceforge.net/projects/tdm-gcc/files/TDM-GCC%20Installer/Previous/1.1309.0/)
• for MATLAB R2017b and beyond: MinGW GCC 6.3 from mingw-w64.org

Instructions

In MATLAB, config the mex compiler using the following commands:

setenv('MW_MINGW64_LOC',<MinGW installation folder>)
mex -setup C++

Edit & Run The Program

In MATLAB, Home -> Open -> 'path-to-repository'/main.c, then click Run on the top.

Project Structure

• /afbs-kernel: the scheduler for Simulink
• /analysis: scripts for analysis results and ploting figures in reports
• /experiments: early experiments
• /result: folder for saving experiment result data
• /tasksets: everything related to taksets
• /tasksets/UUnifast.m: UUnifast synthetic task generator
• main_ga.m: implementations of genetic algorithms
• main_random.m: implementations of random search
• main_cost_space.m: evaluate the cost space
• main_single_period_brute.m: solve the optimal single period using brute force
• main_task_allocation_greedy.m: allocation additional tasks to a non-control taskset
• run_single_simulation: single run of one instance
• rta.py: response time analysis of dual period (in Python)
• .slx: Simulink model files

Known Issues

• EDF policy is not implemented
• There is a limitation on the simulation time (as kernel_cnt is long)
• For DUAL, control tasks have to be the highest priorities
• For DUAL, the deadline is not equal to DUAL
• For DUAL, the switching back too T_i^H is not implemented
• For DUAL, the switch will be deferred to the next release

Citation

@article{dai2019dual,
  title={A dual-mode strategy for performance-maximisation and resource-efficient CPS design},
  author={Dai, Xiaotian and Chang, Wanli and Zhao, Shuai and Burns, Alan},
  journal={ACM Transactions on Embedded Computing Systems (TECS)},
  volume={18},
  number={5s},
  pages={1--20},
  year={2019},
  publisher={ACM New York, NY, USA}
}