Co-evolutionary reinforcement learning for multi-agent systems

Figure 1.
Figure 1.

Objective: To develop co-evolutionary algorithms for reinforcement learning in multi-agent systems

PIs: Chew Lock-Yue (NTU), Bo An (NTU), Mikael Johasson (KTH)

  • Demonstrate the techniques on adaptive scheduling of an intelligent autonomous bus network at NTU campus

Preliminary results: 

  • Proof of concept for elevator scheduling (Figure 1)
  • Data analytics on campus bus network (Figure 2)
  • Preliminary results in simulations (Figure 3)

 

 

 

Figure 2. At peak periods, bus bunching persists due to strong coupling between consecutive buses from loading/unloading of commuters
Figure 2. At peak periods, bus bunching persists due to strong coupling between consecutive buses from loading/unloading of commuters

 

 

 

Figure 3
Figure 3