Combined multiple problem in Scheduling #1252

giangstrider · 2019-05-07T04:44:46Z

giangstrider
May 7, 2019

Hi, I went through many examples of the scheduling problem, but it seems my problem combines many complicated things, and so hard for the newbie like me to do it. If anyone can help or give some suggestion to do it, I'm very appreciated.

I have the following problem:
The main goal is scheduling workers to do the tasks in one week. (Monday -> Friday, 6AM to 6PM per day)

There is a list of tasks and workers.
Workers need to assigned to do the tasks. Each task has a deadline date to do, a duration (unit time to be done)
Each worker has there owned blocking time (to do a meeting for example) in a week, and their blocking time needs to be scheduled an exact date in the week. (It's possible to fix this to start in 6AM before to do another task.). Each blocking time also a duration (unit time to be done).
Each worker has their own skills. Each task required skill for a worker to done the task.
each task has own location. There is a matrix of traveling time between the tasks:
like
A | B
A 0 | 2
B 2 | 0
(A,B is the task)
(Each number is unit of time)

Constraint:

Tasks and Blocking time not overlap.
Each task only assigned to 1 worker.
Task's required skill == worker skills.

Objective:
Minimize traveling time of workers.

The way I modeling is:

Create a list of NewIntervalVar of worker's blocking_time.
Create a list of NewOptionalIntervalVar of tasks. (Since tasks can happen or not).
Add Nonoverlap of [NewIntervalVar_blockingtime, NewOptionalIntervalVar_tasks].

My problem is:

I don't know how to link between the weekday ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday"] within the tasks and blocking_times. (Mean which task assigned to which weekday).
How to add constraint like if the tasks assigned to Monday (for example), if assigned one more task has a long duration, so the worker to need work until 7PM, So this task need shift to Tuesday.
How to compute the objective (traveling time) to minimize.

Feel free to comment or modify my model since I'm not sure it corrects. Thanks very much.

My example code:

all_tasks = []
for block in data['blocked_times']:
    start_blocking = model.NewIntVar(0, horizon, 'start_block_%s_%s' % (block['worker_id'], block['blocked_id']))
    end_blocking = model.NewIntVar(0, horizon, 'end_block_%s_%s' % (block['worker_id'], block['blocked_id']))
    duration = block['activity_duration']
    interval_blocking = model.NewIntervalVar(start_blocking, duration, end_blocking, 'interval_block_%s_%s' % (block['worker_id'], block['blocked_id']))
    all_tasks.append(interval_blocking)


for task in data['tasks']:
    start_task = model.NewIntVar(0, horizon, 'start_task_%s' % (task['task_id']))
    end_task = model.NewIntVar(0, horizon, 'end_task_%s' % (task['task_id']))
    duration = block['activity_duration']
    presence_task = model.NewBoolVar('presence_%s' % task['task_id'])
    interval_task = model.NewOptionalIntervalVar(
        start_task, duration, end_task, presence_task, 'interval_task_%s' % task['task_id']
    )
    all_tasks.append(interval_task)

model.AddNoOverlap(all_tasks)

Answered by lperron

May 11, 2019

You need to map hours to integer values. there are 168 hours in a week. So for one week, the horizon is 168. To force one task to be done on Monday, restrict the end date to be before 24 (assuming Monday is the first day of the week). For the transit, look at example https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ <[email protected]> *Date : *mar. 7 mai 2019 à 06:44 *À : *google/or-tools *Cc : *Subscribed Hi, I went through many examples of the scheduling problem, but it seems my

View full answer

lperron · 2019-05-11T12:30:30Z

lperron
May 11, 2019
Maintainer

You need to map hours to integer values. there are 168 hours in a week. So for one week, the horizon is 168. To force one task to be done on Monday, restrict the end date to be before 24 (assuming Monday is the first day of the week). For the transit, look at example https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ <[email protected]> *Date : *mar. 7 mai 2019 à 06:44 *À : *google/or-tools *Cc : *Subscribed Hi, I went through many examples of the scheduling problem, but it seems my

…

problem combines many complicated things, and so hard for the newbie like me to do it. If anyone can help or give some instruction to do it, I'm very appreciated. I have the following problem: The main goal is scheduling workers to do the tasks in one week. (Monday -> Friday, 6AM to 6PM per day) 1. There is a list of tasks and workers. Workers need to assigned to do the tasks. Each task has a deadline date to do, a duration (unit time to be done) 2. Each worker has there owned blocking time (to do a meeting for example) in a week, and their blocking time needs to be scheduled an exact date in the week. (It's possible to fix this to start in 6AM before to do another task.). Each blocking time also a duration (unit time to be done). 3. Each worker has their own skills. Each task required skill for a worker to done the task. 4. each task has own location. There is a matrix of traveling time between the tasks: like A | B A 0 | 2 B 2 | 0 (A,B is the task) (Each number is unit of time) Constraint: 1. Tasks and Blocking time not overlap. 2. Each task only assigned to 1 worker. 3. Task's required skill == worker skills. Objective: Minize traveling time of workers. The way I modeling is: 1. Create a list of NewIntervalVar of worker's blocking_time. 2. Create a list of NewOptionalIntervalVar of tasks. (Since tasks can happen or not). 3. Add Nonoverlap of [NewIntervalVar_blockingtime, NewOptionalIntervalVar_tasks]. My problem is: 1. I don't know how to link between the weekday ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday"] within the tasks and blocking_times. (Mean which task assigned to which weekday). 2. How to add constraint like if the tasks assigned to Monday (for example), if assigned one more task has a long duration, so the worker to need work until 7PM, So this task need shift to Tuesday. 3. How to compute the objective (traveling time) to minize. Feel free to comment or modify my model since I'm not sure it corrects. Thanks very much. My example code: all_tasks = []for block in data['blocked_times']: start_blocking = model.NewIntVar(0, horizon, 'start_block_%s_%s' % (block['worker_id'], block['blocked_id'])) end_blocking = model.NewIntVar(0, horizon, 'end_block_%s_%s' % (block['worker_id'], block['blocked_id'])) duration = block['activity_duration'] interval_blocking = model.NewIntervalVar(start_blocking, duration, end_blocking, 'interval_block_%s_%s' % (block['worker_id'], block['blocked_id'])) all_tasks.append(interval_blocking) for task in data['tasks']: start_task = model.NewIntVar(0, horizon, 'start_task_%s' % (task['task_id'])) end_task = model.NewIntVar(0, horizon, 'end_task_%s' % (task['task_id'])) duration = block['activity_duration'] presence_task = model.NewBoolVar('presence_%s' % task['task_id']) interval_task = model.NewOptionalIntervalVar( start_task, duration, end_task, presence_task, 'interval_task_%s' % task['task_id'] ) all_tasks.append(interval_task) model.AddNoOverlap(all_tasks) — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <https://github.com/google/or-tools/issues/1252>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACUPL3L7YSPAWPYI3K2Q2ZTPUECMLANCNFSM4HLFJYLQ> .

0 replies

giangstrider · 2019-05-13T07:25:25Z

giangstrider
May 13, 2019
Author

Hi @lperron , thanks for your reply. I'm can model work for the basic case now.
I'm stuck at constraint how to for 1 task is executed by only one worker.
Since the list of tasks is represented by the list of OptionalIntervalVar, with literal BoolVar controlled it.
I don't know how to use BoolVar to model this constraint.

0 replies

lperron · 2019-05-13T07:57:59Z

lperron
May 13, 2019
Maintainer

sum(boolvars) == 1 ? Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ <[email protected]> *Date : *lun. 13 mai 2019 à 09:25 *À : *google/or-tools *Cc : *Laurent Perron, Mention Hi @lperron <https://github.com/lperron> , thanks for your reply. I'm can

…

model work for the basic case now. I'm stuck at constraint how to for 1 task is executed by only one worker. Since the list of tasks is represented by the list of OptionalIntervalVar, with literal BoolVar controlled it. I don't know how to use BoolVar to model this constraint. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/google/or-tools/issues/1252#issuecomment-491706370>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACUPL3NI2UQZAEXXW3NGGNLPVEJW7ANCNFSM4HLFJYLQ> .

0 replies

giangstrider · 2019-05-13T09:07:46Z

giangstrider
May 13, 2019
Author

Thanks @lperron , I use same methods but the wrong way. It's ok now.
For the transit, at this line:
https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py#L104

I'm wondering how to identify the distance between two jobs. For example in jobshop_ft06_distance_sat, https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py#L34
How j2 - j1 can identify as distance?
How can I get input my custom data?

Is there any way to access the metadata of the job which defined before? Like j1.name? (From name I can extract the ID and mapping to my custom data)

0 replies

lperron · 2019-05-13T09:54:11Z

lperron
May 13, 2019
Maintainer

Because in this toy problem, distance(j1, j2) == abs(j2 - j1). Please replace with any custom distance. Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ <[email protected]> *Date : *lun. 13 mai 2019 à 11:07 *À : *google/or-tools *Cc : *Laurent Perron, Mention Thanks @lperron <https://github.com/lperron> , I use same methods but the

…

wrong way. It's ok now. For the transit, at this line: https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py#L104 I'm wondering how to identify the distance between two jobs. For example in jobshop_ft06_distance_sat, https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py#L34 How j2 - j1 can identify as distance? How can I get input my custom data? Is there any way to access the metadata of the job which defined before? Like j1.name? (From name I can extract the ID and mapping to my custom data) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/google/or-tools/issues/1252#issuecomment-491740771>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACUPL3OHPE62FUPGHT4S653PVEVWVANCNFSM4HLFJYLQ> .

0 replies

lperron · 2019-05-13T09:55:26Z

lperron
May 13, 2019
Maintainer

About metadata, I use the named tuples for that. Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *Laurent Perron <[email protected]> *Date : *lun. 13 mai 2019 à 11:53 *À : *google/or-tools *Cc : *google/or-tools, Mention Because in this toy problem, distance(j1, j2) == abs(j2 - j1).

…

Please replace with any custom distance. Laurent Perron | Operations Research | ***@***.*** | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ ***@***.***> *Date : *lun. 13 mai 2019 à 11:07 *À : *google/or-tools *Cc : *Laurent Perron, Mention Thanks @lperron <https://github.com/lperron> , I use same methods but the > wrong way. It's ok now. > For the transit, at this line: > > https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py#L104 > > I'm wondering how to identify the distance between two jobs. For example > in jobshop_ft06_distance_sat, > https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py#L34 > How j2 - j1 can identify as distance? > How can I get input my custom data? > > Is there any way to access the metadata of the job which defined before? > Like j1.name? (From name I can extract the ID and mapping to my custom > data) > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <https://github.com/google/or-tools/issues/1252#issuecomment-491740771>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ACUPL3OHPE62FUPGHT4S653PVEVWVANCNFSM4HLFJYLQ> > . >

0 replies

giangstrider · 2019-05-13T13:42:30Z

giangstrider
May 13, 2019
Author

Thanks for your help @lperron . I able to make transit work.
just a little extend about the transit,
The jobshop distance example works for the transit between tasks, but how about transit between from the beginning to the task location?
From start of the day, there is a distance moving from the worker's home to the task's location = X
So for example, the task start of interval var is 24 (beginning of Tuesday), I would like to add moving time, so the start will be: start = start + X.

I'm thinking modify somewhere in this for loop https://github.com/google/or-tools/blob/master/examples/python/jobshop_ft06_distance_sat.py#L88
But unable to know how to identify it the solver makes the task start first in the day to add the moving.

0 replies

lperron · 2019-05-13T14:16:49Z

lperron
May 13, 2019
Maintainer

Same thing, you have one boolean from the dummy node to each task. You can add a enforced linear equation using this boolean. Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ <[email protected]> *Date : *lun. 13 mai 2019 à 15:42 *À : *google/or-tools *Cc : *Laurent Perron, Mention Thanks for your help @lperron <https://github.com/lperron> . I able to make

…

transit work. just a little extend about the transit, The jobshop distance example works for the transit between tasks, but how about transit between from the beginning to the task location? From start of the day, there is a distance moving from the worker's home to the task's location = X So for example, the task start of interval var is 24 (beginning of Tuesday), I would like to add moving time, so the start will be: start = start + X. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/google/or-tools/issues/1252?email_source=notifications&email_token=ACUPL3PGJP6MY4RNVSSQSC3PVFV47A5CNFSM4HLFJYL2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVILBXA#issuecomment-491827420>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACUPL3MF7VX3DJ5VRJAVQQLPVFV47ANCNFSM4HLFJYLQ> .

0 replies

giangstrider · 2019-05-14T03:36:33Z

giangstrider
May 14, 2019
Author

Thanks a lot! @lperron
I'm struggling with next constraint which shifts the task if span to next day.
You have a suggestion before To force one task to be done on Monday, restrict the end date to be before 24 (assuming Monday is the first day of the week).
But how to model it in general?

For example, If the task starts at 47 (Tuesday) and has a duration = 3.
This task will end at 50, it's mean span to Wednesday.
Because I don't know the solver will scheduling in which start time, so I don't know the value to add constraint like model.Add(end < some_value)
How can I prevent this, moving this task to another time slot if the task span from today to the nextday?

Seconds, there is 2 objective I would like to Minimize:

Minimize Transit time.
Which variable I can get to get sum(transit_time) to calculate in jobshop_ft06_distance_sat?
Task location change.
It's better to minimize the times of location change between the tasks.
I have no idea to calculate this objective. Can you suggest for me?

Thanks

0 replies

lperron · 2019-05-14T06:01:41Z

lperron
May 14, 2019
Maintainer

You need to restrict the start time to only valid values. For this, use a special domain for the start var. Please note that the API to use this has changed with or-tools 7.1. Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ <[email protected]> *Date : *mar. 14 mai 2019 à 05:36 *À : *google/or-tools *Cc : *Laurent Perron, Mention Thanks a lot! @lperron <https://github.com/lperron>

…

I'm struggling with next constraint. For example, If the task starts at 47 (Tuesday) and has a duration = 3. This task will end at 50, it's mean span to Wednesday. How can I prevent this, moving this task to another time slot if the task span from today to the nextday? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/google/or-tools/issues/1252?email_source=notifications&email_token=ACUPL3OVHPSEBLMZAQNDJKDPVIXUXA5CNFSM4HLFJYL2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVKFMFY#issuecomment-492066327>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACUPL3KQI23CFDWQXK7JIHTPVIXUXANCNFSM4HLFJYLQ> .

0 replies

giangstrider · 2019-05-18T15:20:27Z

giangstrider
May 18, 2019
Author

Hi @lperron , I've back from sick, thanks for your help.
After experienced a few days, I found my model not correct for this problem.

just a little extend about the transit,
The jobshop distance example works for the transit between tasks, but how about transit between from the beginning to the task location?
From start of the day, there is a distance moving from the worker's home to the task's location = X
So for example, the task start of interval var is 24 (beginning of Tuesday), I would like to add moving time, so the start will be: start = start + X.

Same thing, you have one boolean from the dummy node to each task. You can add a enforced linear equation using this boolean. Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00
…

As I understand, For example, we have 2 workers and 3 tasks.
That's mean for each worker, we create 3 start_literal for 3 tasks, and only one start_literal=True.
In my problem, for each Monday, Tuesday ... til Friday, look like each day need 1 start_literal=True (5 start_literal = True for the whole week),
because each day, the worker will move from home to the task's location. Is there any way or example to model this?
Next, in my code I model this line:
model.Add(starts_assignment[idx_i] >= worker_to_location_distance).OnlyEnforceIf(start_literal)
I make constraint starts_assignment[idx_i] >= worker_to_location_distance, but it's not correct.
From dummy node (For each day), the start of node need >= distance + X with X is the start_int_var. But i don't know how to get start_int_var, because it depends on the task happens on which weekday.
Maybe there is anyway model can solve both problems? Because I think it related.

Here is my code

arcs = []
    for idx_i in range(len(task_intervals)):
        # dummy node
        start_literal = model.NewBoolVar('%i_first_job' % idx_i)
        end_literal = model.NewBoolVar('%i_last_job' % idx_i)
        arcs.append([0, idx_i + 1, start_literal])
        arcs.append([idx_i + 1, 0, end_literal])
        i_point = idx_i

         # Worker's home to task's location model
        worker_point = workers[idx_i]
        worker_to_location_distance = toy_distance_1
        model.Add(starts_assignment[idx_i] >= worker_to_location_distance).OnlyEnforceIf(start_literal)

        for idx_j in range(len(task_intervals)):
            if idx_i == idx_j:
                continue

            literal = model.NewBoolVar('%i_follows_%i' % (idx_j, idx_i))
            arcs.append([idx_i + 1, idx_j + 1, literal])

            j_point = idx_j
            model.Add(starts_tasks[idx_j] >= ends_tasks[idx_i] + toy_distance_2).OnlyEnforceIf(literal)

    model.AddCircuit(arcs)

0 replies

lperron · 2019-05-18T21:21:08Z

lperron
May 18, 2019
Maintainer

I would imagine creating a dummy node and each night break, and adding it in the circuit. I think you need 1 circuit per person. Laurent Perron | Operations Research | [email protected] | (33) 1 42 68 53 00 *De : *☂ Giang Strider ☂ <[email protected]> *Date : *sam. 18 mai 2019 à 17:20 *À : *google/or-tools *Cc : *Laurent Perron, Mention Hi @lperron <https://github.com/lperron> , I've back from sick, thanks for

…

your help. After experienced a few days, I found my model not correct for this problem. just a little extend about the transit, The jobshop distance example works for the transit between tasks, but how about transit between from the beginning to the task location? From start of the day, there is a distance moving from the worker's home to the task's location = X So for example, the task start of interval var is 24 (beginning of Tuesday), I would like to add moving time, so the start will be: start = start + X. Same thing, you have one boolean from the dummy node to each task. You can add a enforced linear equation using this boolean. Laurent Perron | Operations Research | ***@***.*** | (33) 1 42 68 53 00 … <#m_7971818429051772032_> As I understand, For example, we have 2 workers and 3 tasks. That's mean for each worker, we create 3 start_literal for 3 tasks, and only one start_literal=True. In my problem, for each Monday, Tuesday ... til Friday, look like each day need 1 start_literal=True (5 start_literal = True for the whole week), because each day, the worker will move from home to the task's location. Is there any way or example to model this? Here is my code arcs = [] for idx_i in range(len(task_intervals)): # dummy node start_literal = model.NewBoolVar('%i_first_job' % idx_i) end_literal = model.NewBoolVar('%i_last_job' % idx_i) arcs.append([0, idx_i + 1, start_literal]) arcs.append([idx_i + 1, 0, end_literal]) i_point = idx_i for idx_j in range(len(task_intervals)): if idx_i == idx_j: continue literal = model.NewBoolVar('%i_follows_%i' % (idx_j, idx_i)) arcs.append([idx_i + 1, idx_j + 1, literal]) j_point = idx_j model.Add(starts_tasks[idx_j] >= ends_tasks[idx_i] + toy_distance).OnlyEnforceIf(literal) model.AddCircuit(arcs) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/google/or-tools/issues/1252?email_source=notifications&email_token=ACUPL3OZJ2WMRY5T73POE2DPWANEHA5CNFSM4HLFJYL2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVWQQUQ#issuecomment-493684818>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACUPL3JHMUW6WN4V7IQATGDPWANEHANCNFSM4HLFJYLQ> .

0 replies

giangstrider · 2019-05-20T09:29:56Z

giangstrider
May 20, 2019
Author

Hi @lperron , can you more clarify, how I can define a night break.
And with N workers, it's mean creating N * 5 (Monday..Friday) dummy node per worker, and add it in the circuit? (So N times add to the circuit?)
Am I right?

0 replies

lperron · 2019-05-20T10:03:47Z

lperron
May 20, 2019
Maintainer

You will need 1 circuit per worker. 5 extra night nodes per circuit. Then 1 constraint per day shift that says 1 shift is done by exactly 1 worker. Le lun. 20 mai 2019 à 11:30, ☂ Giang Strider ☂ <[email protected]> a écrit :

…

Hi @lperron <https://github.com/lperron> , can you more clarify, how I can define a night break. And with N workers, it's mean creating N * 5 (Monday..Friday) dummy node per worker, and add it in the circuit? (So N times add to the circuit?) Am I right? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/google/or-tools/issues/1252?email_source=notifications&email_token=ACUPL3JQDIXF4IZIHNMGCYDPWJVR5A5CNFSM4HLFJYL2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGODVYHM3Y#issuecomment-493909615>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ACUPL3NP7GEOKZCQDK7QAN3PWJVR5ANCNFSM4HLFJYLQ> .

0 replies

giangstrider · 2019-05-22T03:56:32Z

giangstrider
May 22, 2019
Author

Sorry @lperron , I still stuck with this.
Here is sample code I model follow my understanding, can you take a look?
For each worker,
for each task, I added 5 night_break_node, but not sure the source node and destination of this night break node will be?
(Can I send you full code in private, to run the test if you need, just my thinking)

weekdays = range(0,5)
workers = range(0,3)
tasks_starts = [1_start, ..., 5_start]
tasks_intervals = [1_interval, ..., 5_interval]
tasks_ends = [1_end, ..., 5_end]

for worker in workers:
    arcs = []


    for t1 in range(len(tasks_intervals)):
        start_lit = model.NewBoolVar('%i_start_node' % (t1))
        arcs.append([0, t1 + 1, start_lit])
        
        w_to_t1_distance = toy_distance(w, t1)
        
        # 5 extra night nodes
        for w in weekdays:
            night_break_node_literal = model.NewBoolVar('%i_night_break_%i' % (w, t1))
            # Add night break node
            arcs.append([t1 + 1, 0, night_break_node_literal])

            task_start_bound, task_end_bound = toy_bound(w) # (12,24) is bound limit of w=1(Tuesday) for example
            model.Add(tasks_starts[t1] >= task_start_bound + inspector_to_i_distance).OnlyEnforceIf(start_lit)
            model.Add(tasks_ends[t1] <= task_end_bound)

        for t2 in range(len(tasks_intervals)):
            if t1 == t2:
                continue

            lit = model.NewBoolVar('%i_follows_%i' % (t1, t2))
            arcs.append([t1 + 1, t2 + 1, lit])

            t1_to_t2_distance = toy_distance(t1, t2)
            model.Add(tasks_starts[t2] >= tasks_ends[t1] + t1_to_t2_distance).OnlyEnforceIf(lit)

    model.AddCircuit(arcs)

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Combined multiple problem in Scheduling #1252

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Replies: 15 comments

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

{{title}}

{{title}}

{{title}}

{{title}}

{{editor}}'s edit

{{editor}}'s edit

Select a reply

Combined multiple problem in Scheduling #1252

giangstrider May 7, 2019

Replies: 15 comments

lperron May 11, 2019 Maintainer

giangstrider May 13, 2019 Author

lperron May 13, 2019 Maintainer

giangstrider May 13, 2019 Author

lperron May 13, 2019 Maintainer

lperron May 13, 2019 Maintainer

giangstrider May 13, 2019 Author

lperron May 13, 2019 Maintainer

giangstrider May 14, 2019 Author

lperron May 14, 2019 Maintainer

giangstrider May 18, 2019 Author

lperron May 18, 2019 Maintainer

giangstrider May 20, 2019 Author

lperron May 20, 2019 Maintainer

giangstrider May 22, 2019 Author

giangstrider
May 7, 2019

lperron
May 11, 2019
Maintainer

giangstrider
May 13, 2019
Author

lperron
May 13, 2019
Maintainer

giangstrider
May 13, 2019
Author

lperron
May 13, 2019
Maintainer

lperron
May 13, 2019
Maintainer

giangstrider
May 13, 2019
Author

lperron
May 13, 2019
Maintainer

giangstrider
May 14, 2019
Author

lperron
May 14, 2019
Maintainer

giangstrider
May 18, 2019
Author

lperron
May 18, 2019
Maintainer

giangstrider
May 20, 2019
Author

lperron
May 20, 2019
Maintainer

giangstrider
May 22, 2019
Author