Page 1 of 1

Assigning public mental health workers to new clients

Posted: Sat Jun 11, 2016 9:21 pm
by philtroy

I wish to simulate the assignment of mental health workers in a public health care system to new clients subject to the following constraints:

- The mental health worker needs to respond to the new client within a certain number of days (usually 5 working days)

- The first meeting between a mental health worker and a client typically is longer than other meetings (90 minutes)

- The mental health worker needs to write up his/her notes within 2 days of each meeting with the client

- The number of meetings each client has with the mental health worker they see first follows a probability distribution

- Subsequent meetings should not occur before one week after the most immediate previous meeting

- Subsequent meetings should occur before two weeks after the most immediate previous meeting

- Mental health workers take vacations, at pre-scheduled (non-random times). Thus it is important to not assign a mental health
worker that is about to go on a two week vacation to a new client unless the mental health worker has enough time to see the client,
plan the follow-up and write-up notes before starting the vacation (while still meeting other time deadlines)

All the probability distributions (number of new clients per day, number and type of visits per client) are not yet known but will be estimated. The overall challenge here is to determine how to assign clients to the mental health workers so as to ensure that violations of constraints are minimized or precluded and so as to also minimize the number of mental health workers needed to serve the clients.

I have some thoughts as to how to go about this, but, they involve a lot of visual logic (building a list of activities for each worker, building a calendar for each worker, assigning tasks to the calendar, . . .). But since that approach will end up taking a fair amount of time, I am interested in any ideas any of you have to take advantage of the basic building blocks of Simul8 so that I can implement this model as quickly as possible.

Thanks . . .

Phil Troy

Philip M. Troy, Ph.D.

Senior Analytics Advisor
West Central Montreal Integrated Uniiversity Health And Social Services Network

professeur adjoint de chirurgie
Université McGill