Model of QAP Footprints Training

Results

rl-uni-20

• use dataset with qap instance of size 20
• number of footprints for the training : 2000
• success of prediction after training : 73,675%

rl-uni-30

• use dataset with qap instance of size 30
• number of footprints for the training : 2000
• success of prediction after training : 76.05%

rl-uni-30-V2

• use dataset with qap instance of size 30 generated with random walk v2 to generate square images
• number of footprints for the training : 2000
• success of prediction after training : 78.55%

rl-uni-20&30

• use dataset with qap instance of size 20 and 30
• number of footprints for the training : 4000
• success of prediction after training : 73.1875%