(Fig. 1 Network of BASIS Phoenix)

The plan is to now generate a demand file, where an estimation of the total travel demand is calculated. Then, we can create a .yaml file storing the settings of the simulation including constraints. And finally, run it, which will produce the two agent.csv and trajectory.csv files. The simulation is quite complicated so I won’t go much into detail here. But the general idea is that within each physical lane, exists several “micro-links” where cars can change lanes in. This adds an additional level of complexity and interest, as we’re able to see the individual car’s reactions to splitting the lane for different use cases. In the future, my idea is to compare the smart reservation system’s outputs as well as the normal unchanged status quo, to see if it is truly worth it to switch over to a new system. For the probability of switching over, I’ll be using a logit model, a probability model that measures the likelihood of individuals switching over to the new system, something I covered in the last blog post.

Today, I’ll be working on developing the things mentioned. Hopefully, by next week we can have some sort of functioning code that will run the status quo as of right now.

That’s all we have for this week. Thanks for reading this post, and I’m going to get coding!