Open thomasGodfrey opened 3 years ago
Cool feature. What of this needs exposing in the modelling language (what might differ between EDs?)?
In the modeling language, I would consider:
A table structure whereby each row is a patient and each column is some information about that patient.
The user could add columns as they wish
Each column would effectively form a category of 'belief' that the agent may hold...
The language would have to explicitly expose what entity the column concerns (patient, staff, resource) so that the beliefs can be categorised appropriately
We may also consider that the user can specify custom orderings of the waiting list itself (e.g. by severity, by patient LoS). This would influence, but not override the staff's patient selection. Effectively acting as giving a column's belief more weight.
This is a work in progress...
But as we hopefully move to an agent implementation closer to BDI, these beliefs could then be considered in agent decision-making
Good start. Remember to differentiate between simulation state and the modelling of processes and tools. Your table above sounds a bit like it's modelling simulation state; in particular, you're talking about specific patients. I wonder whether the more relevant information at the DSML-level is about the schema (i.e., the column headers of the table without the rows).
I would also expect something in the model about how agents interact with the table. How do they add things, how do things get updated or removed, how and when do agents learn about changes in the table, ...
Implement a waiting list for patients based on my observations in the ED.