This content has been marked as final. Show 4 replies
There is no perfect solution for your problem with the current solver. But you can still solve it with the following workaround. Assuming you are expecting to visit all orders in a week, the idea is to create a long continous route per driver for the whole week, assign orders into it and then break the weekly route into 5 daily routes afterwards. Here are the steps to setup the problem:
Create 4 routes, one for each driver. Set the route start time to 8am, set the MaxOrderCount to 150 (or even larger), leave MaxTotalTime, MaxTotalTravelTime and MaxTotalDistance field empty. Set the route end depot to the specified depot.
For each route, introduce 5 MaxCumulWorkTime type of breaks with the proper MaxCumulWorkTime value (i.e. 4hrs for the first break, 13hrs for the second break if the lunch time is 30mins long.)
Solve the VRP layer and break each route into 5 shorter routes.
Another option is to create a route for a driver on each day of the week. See the post here:
You can use the time-window break to represent the lunch break in this case. The only problem with this idea is you have to set at least one depot (either start or end depot) for a route. You might use some hotel location as the end depot in the middle of the week.
Hope this is helpful.
I am not sure if this will work, but you could scale down you 7 day analysis to 24 hours. that is 0th hour on day 1 will be 0th hour on your 24 hour clock and the 24th hour of your 7th day will be 24th hour on the 24 hour clock. Similarly you can define lunch time and weekends on the same scale. Later when you get your answer, you can scale them up again. For display and mapping purposes you could use a evaluator field that gives your the correct real time from the scaled time.
I never did it, but just an idea.
Overview of my project.
I have 600 sampling locations and 4 vehicles (150 stops per driver). The locations can only be visited between 8 am and 4:30 pm. Drivers have 30 minute break in middle of the day.
What i would like to accomplish are routes that visit the sampling location during time window then assume the driver continues toward next location, stays overnight and starts with next set of points the following day. I want it to repeat this until all 150 points are visited. I would also like routes to return to a specified depot during weekends if possible.
I have ran several attempts at this but keep getting violated constraints. It seems as though network analyst doesn't like that routes continue past the Specific Date that I set as a start date in the Analysis Settings.
Has anyone successfully created routes spanning multiple days with time constraints? Any help would be appreciated.
I have a similar / same issue.
I tried Anna's suggestion of setting 5 MaxCumulWorkTime breaks.
However, as I have a TimeWindowStart and TimeWindowEnd set, the solve stops at the end of the first day.
What has developed for planning multiple day routes since the orignal post was made?
Longer version of my situation:
- 49 sites across the state
- 3hr inspection per site
- site only open 8am to 6pm
- must arrive before 3pm
- can be more than four hours drive to next site
- 3 inspectors
- guestimate six people weeks of work
- semi flexible on start / end of week; ie can drive to first inspection Sunday so can start 8am Monday to make second site on day one
- able to trial flying to different start points to see quicker / cheaper option; thee main airports in state
- bonus would be where should the driver sleep that night
- stay at last site
- drive to tomorrow morning site
- drive part way to tomorrow morning site
- max hours allowed to work / drive per day limited
As 'only' 49 sites I have made five routes ignoring any constraints and can manually adjust.
Thanks for your time