Addition of Vehicle Scheduling Program

[Raghucharan16]

Having multiple buses with multiple routes with different trip times in an individual day itself is normal, with the limitation of start time of bus with in an interval, and scheduling the buses effectively based on occupancy demand and buses should not be idle for very long time would be a great needed scheduling problem. Needed help while implementing this problem.

[TomCools]

Could you please clarify how you want to divide this problem? What things are missing now VS what is the new problem you want to solve? Depending on how you specify, the solution could be very different, even ranging form 1 solution, to 2 or 3 solutions which need to be combined.

• Scheduling effectively based on occupany mostly says something about a Type of Bus (large vs small)
• "busses should not be idle too long" seem to refer to breaks between 'runs'?

[Raghucharan16]

constraints are more of like each bus getting started must complete a return trip in 24hours, assuming if a bus starts at night, it should reach before that time next day. and also schedule should be respective to individual days. i.e ideally for a normal day, there should be 2 trips per bus, but in peak days like weekends, special days, festival days, we have to adjust the timings accordingly so that we may run 3 trips for a bus in order to maximize the occupancy. yes ofcourse it depends on previous trip's arrival but we have to find a way to adjust the timings inorder to maximize the occupancy. this is actually the most crucial part. adjusting the timings and coming to a decision of doing 3 trips and also adjusting the timings of other bus too, because when a bus goes to B from A, then B to A and again A-B, any of buses from B-A should stop at A inorder to make trip rotation in that time not the exact time slot but near by time slot. how can we effectively let the system generate schedule those trips without messing bus availability at the other station, that is, schedule on monday can be different because of occupancy levels and can run only 16 trips [8 buses 2 per bus], but on thursday,friday, saturday there can be 18 trips or 19 trips based on occupancy. but that should be generated by algorithm with ease without effecting the schedule.

[TomCools]

Where does the occupancy come from in your story? Is that known up front or something which changes dynamically during the day?
Do I get you correctly that a 3rd trip is not guaranteed but you expect it to be determined by the scheduler?

[Raghucharan16]

So, we will maintain a csv which contains 0-48 slots representing 48 half hours time interval through out the day. and corresponding to each slot say {00:00-00:29, 00:30-00:59, etc}, we'll have an occupancy value. [lets say it will be of 0-1 range].
And yes you got it right, 3rd trip is not mandatory but if it gives good occupancy, [lets say there is some threshold, to check] then it should schedule a trip for a 3rd trip. and if it is done, next day timings should be taken care as depends on bus availability by the scheduler itself or it can give approximate schedule and in the end, we can look over it and decide the schedule..

[TomCools]

I don't fully grasp your "occupancy" level here. Maybe some basic questions to check with you:

• Do busses always leave on a set time and arrive at a set time? Or can this be different day by day? (In Belgium, the schedule is set and can't just change)
• Is the occupancy based on the amount of people per timeslot? Does this mean the occupancy is only for a single stop, and not all possible stops?

[Raghucharan16]

what do you mean by set time? in our place, [Hyderabad and Bangalore], based on traffic, a trip time may changes in a single day itself.
for example, for early morning trips it may take some 6hours, in the even it may stretch up to 8hrs too, and late night it[trip time] may fall to 7 hours. and coming to occupancy, it too depends on festivals, weekends, some special govt holidays. and i consider occupancy for complete trip. lets say for point A, there may be x boarding points and point B may have y dropping points, but i consider total occupancy as total no.of seats filled irrespective of boarding and dropping points. for more clarity,
i have a csv like this,
timeslot,occupancy
00:00-00:29, 0.03
05:00-05:29,0.6
14:00-14:29,0.47
22:00-22:29,1

i just take highest occupancy slot [basically no.of total passenegers] and divide every other slot with this to make the level in (0-1).
Both the trip time and occupancy levels are dynamic and may subject to change and vary on many factors.
so in a detailed csv, we can have attributes like,
timeslot,route,dayofweek,traveltime,occupancy