Seeking Solutions for common scheduling challenges of finding multiple possible future reservations

[Christian Nielsen]

Hello everyone,

I hope this message finds you well. I am currently tackling two distinct problems, both of which share a common core issue. I am reaching out to see if any of you might have some valuable suggestions to offer.

Case 1: Multiple meeting suggestions for the same meeting
In one scenario, I am tasked with scheduling a series of meetings in advance, involving multiple participants. The challenge lies in identifying possible dates, considering a variety of internal restrictions. Optaplanner has proven effective in handling this, generating potential date-series. However, a potential hiccup emerges when participants might not be available on these generated dates. To cater to this, we need to present participants with a selection of more possible dates than needed. Let's say, six suggested dates that will be tentatively reserved in the system until we receive their responses. From there, participants finalize  only three meeting dates, while the remaining  three dates must be released for other potential meetings.

My current approach is a bit straightforward – seeking six dates for a three-day meeting, although I know only three days are needed. While this approach works, ensuring all constraints are met becomes challenging. For example, meetings must occur consecutively, and certain combinations of the six dates are valid while others aren't. I've been contemplating a solution involving modelling three distinct sets of identical meetings, each set adhering to internal rules. 

However, this approach inadvertently collide with  the overlapping rules-restrictions  so it would produce 3 sets of meetings that does not share any dates among them. And if I remove that rules for those particular suggestion, then all 3 sets often becomes identical and hence not 3 real differnet options. I'm unsure how to address this particular issue or if its the right path. maybee som special rule to encure the sets does differ somehow, could work.

I believe this is a challenge that others might have encountered as well. If anyone has insights or solutions to share, I would greatly appreciate it.

Case 2: Vehicle Routing and ad hoc addition of pickup 

In another similar scenario related to vehicle routing, I am tasked with managing routes involving pickups and dropoffs, while the vehicles are allready out delivering. The twist here is that I need to offer customers a variety of pickup/dropoff timeslots to choose from, before they finalize their orders. This situation echoes the same core issue as the previous case, albeit with a different context.

Consider that I have three timeslots available each day, adding up to 21 slots weekly. The challenge here is to figure out all possible pickup/dropoff timeslots for a given customer's location and present them with a selection of options, real-time as the vehiclea are allready out delivering the orders they know. My current approach involves inserting dummy reservations for each customers location in each timeslot, checking if a viable solution exists in each. This process results in 21 full solutions for a single week, for a single customer request, which is quite time-consuming.

I have a feeling that this challenge might be a common one in various contexts. If anyone has encountered similar scenarios or has innovative solutions to suggest, I am all ears.

Thank you for taking the time to read my post. Your insights could greatly help in addressing these intricate scheduling issues. Looking forward to your suggestions and thoughts.

Best regards,