Flowtime single resource

[tpaviot]

Related to issue #50

to use:

problem.add_objective_flowtime_single_resource(worker_1, time_interval = [40, 50])

[dreinon]

Would this already work for time intervals :open_mouth: ?

[codecov-commenter]

Codecov Report

Merging #56 (33a14ff) into master (ceeea8a) will decrease coverage by 0.05%. The diff coverage is 92.94%.

@@            Coverage Diff             @@
##           master      #56      +/-   ##
==========================================
- Coverage   94.86%   94.81%   -0.06%     
==========================================
  Files          30       30              
  Lines        3272     3354      +82     
==========================================
+ Hits         3104     3180      +76     
- Misses        168      174       +6     

Impacted Files	Coverage Δ
test/test_indicator.py	95.62% <89.83%> (-3.41%)	:arrow_down:
processscheduler/problem.py	97.03% <100.00%> (+0.67%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update ceeea8a...33a14ff. Read the comment docs.

[tpaviot]

Currently only for one single time interval. You cannot pass a list of time intervals. And I did not test to add multiple single intervals, for example:

problem.add_objective_flowtime_single_resource(worker_1, time_interval = [40, 50])
problem.add_objective_flowtime_single_resource(worker_1, time_interval = [60, 70])

[dreinon]

Well still a big progress!! Thanks you!

[tpaviot]

This needs to be tested but I merge the branch. I think this implementation is not efficient in terms of computing time, we'll see later for any improvement, let's first check the expected behavior is correct.

[dreinon]

Since this is finally an objective, its API for many time intervals would be add many objectives right? About what you were saying about performance, I wonder if this volume of objectives would enlarge the problem too much. In my case, I would have to add:

• 1 objective/(time_intervalresource) 7 time_intervals * 20 resources = 140 objectives

I can try it out though and see how it works! Other possibility is to first test it before applying it to my problem, so I could also make some tests.

Also, I wonder if this would currently work in the new multiobjective incremental solver, with other max objective like utilization, since these flowtime_single_resource are min objectives right?

Thank you!

[tpaviot]

Ok, let me know about your tests. I ran a few tests with the new multiobjective incremental solver, it's much more convenient than the previous optimizer.

[dreinon]

Nice! Although with much more convinient, you mean it might be duable to have 140 objectives? And also, is the new solver already compatible with mixed max and min objectives? (to make tests mixing them)

[tpaviot]

Don't know what it gives with so many objectives, you should test (first try with 2, 4, 10, before trying 140!). The new multi-objective solver can deal with both min/max objectives (see last commit fb64fff).

[dreinon]

Yes of course, I'll start simple haha. Thanks!