Single Queue example model in TreeAgePro 2018.
TreeAge Pro 2018 R1.0 introduced new node types (Stop and Entry nodes), which make it easier to build queueing models using fixed cycle Markov structure. Another new feature which is important for queue modeling are Global Trackers, which can be used to store values specific to a particular strategy and shared by all trials (patients) evaluated within that strategy. The attached model implements simple First In First out queue with exponential arrival time and exponential service time. Two strategies can be assigned different numbers of resources (e.g. hospital beds, doctors, etc.) allowing for comparison of maximum queue lengths, resource utilization, etc. under different parameters.
In addition to a TreeAge Pro model, there is a document describing the model in more detail as well as overview of queuing theory.
Comments
Hello Mohamed,
Currently you can only implement the prioritization or any other sophisticated resource constraint logic using Python. We are considering enhancements in the future to allow such resource constraint logic to be expressed within TreeAge Pro framework without using Python, but we have not adequately defined such requirements. One problem is that with Python you can define and describe many different queuing / resource constraint regimes (FIFO, LIFO, priority (with and without pre-emption, etc.). Since Python at least in theory can be used to implement any of them, we struggle with defining something within TreeAge Pro, that may or may not be useful. For us it is purely a question what is the most important set of features for the largest number of users. If we have a subset of requirements for resource constraint implementation that has a narrower scope and meets many users application needs, that would help us see and prioritize such future development.
Could you help us understand a bit more about how you would like to obtain the prioritized list of prognostic/risk scores?(e.g. will this be a sample from a distribution? The model would have to be run in parallel trials mode?
Perhaps other users could add their comments about resource constraint modeling and what features they would like to see to be implemented in "native" TreeAge Pro functionality, rather than resorting to Python.
Thank you for raising this question.
Al
Hi Al,
I'm not familiar with Python programming so was wondering if there are any example Python scripts out there that can execute allocation of resources to trials based on priority; possibly ordering of patients based on prognostic/risk score (e.g AMI/stroke risk) calculated from baseline variables. Is there another way of implementing this prioritisation within Treeage without using Python?
Thanks!
Please sign in to leave a comment.