Easier dataset management, absence-aware shift scheduling, and SLA-aware visit prioritization

Today, we're announcing version v1.10 of the Timefold Platform and updates to the Timefold models.

This new version of the Timefold Platform comes with these platform improvements:

7c6e95e5d532a2519cf70837618a0880190678b6a63f4896064f6f5f543ae0c7

Clearer navigation and faster access to documentation in the new header
: The header shown across every page has been redesigned with a clearer two-band layout. It includes a helpful description and a documentation link that adjusts to the page you're currently viewing.
Manage queued datasets more easily
: You can now manually delete a queued dataset instead of waiting for it to start automatically or expire. The solve queue table also has a new Attempts column, showing how many times a dataset has been attempted. See solve queue for more details.
Re-solve an invalid dataset without resubmitting it
: When a dataset ends up in an invalid state, you can now re-solve it directly from a button, instead of submitting a new dataset from scratch. See invalid state for more details.
See each plan's latest revision on the Plans overview
: A plan's row on the Plans overview now shows the scores and metrics of its latest revision, instead of its original submission. This keeps the overview accurate once a plan has been re-solved or patched, without needing to switch to the Revisions tab. See Plans overview for more details.
Faster solving by pausing automatic score analysis refresh in the UI
: The Platform UI no longer automatically re-fetches score analysis while a dataset is still solving. Doing so repeatedly competed for compute resources with solving itself, which could slow it down. You can still follow how scores and output metrics evolve in real time on the dataset's solve graphs, and if you're calling the score analysis endpoint directly, you can still request it as often as you like while a dataset solves. See score analysis for more details.

Next to that, this new version of the Timefold Platform comes with updates to these Timefold Models:

Field Service Routing (v1 | Stable)

Default priority for visits without an explicit priority
: Previously, a visit that did not specify its own priority always fell back to the built-in priority 6. A new
defaultPriority
field in
config.model.overrides
lets you change this fallback for the entire dataset, without having to set
priority
on every visit individually.
defaultPriority
accepts any built-in priority (
1
-
10
,
opt-1
-
opt-10
) or, when custom priorities are configured, one of your custom priority names. See Default priority for more details.
New constraints to penalize unassigned visits based on their SLA deadline
: Route optimization now includes two new constraints, disabled by default, that penalize visits left unassigned according to how tight their
latestSlaEndTime
is relative to the other visits in the dataset:
Require scheduling mandatory visits with tighter SLA
(medium) penalizes unassigned mandatory visits, more so the tighter their SLA deadline is compared to the rest of the dataset, and
Prefer scheduling optional visits with tighter SLA
(soft) does the same for optional visits. These give you an alternative to the existing priority-based
Require scheduling mandatory visits
and
Prefer scheduling optional visits
constraints for deciding which visits to leave unassigned when there is not enough capacity to schedule them all. See Penalize unassigned visits based on SLA deadline for more details.
Maintenance release
: This release also updates dependencies to deliver security updates.

Employee Shift Scheduling (v1 | Stable)

Absence-aware consecutive days worked (preview)
: Consecutive days worked rules now support the
absenceMode
field, which bridges over days an employee is fully unavailable instead of letting them break the consecutive-days streak. When set to
ABSENCE_AWARE
, two worked days separated only by fully-absent days are treated as consecutive: for example, an employee who works Monday and Tuesday, is fully unavailable Wednesday, and works again Thursday, is treated as having worked four consecutive days, not two separate streaks.
absenceMode
defaults to
ABSENCE_UNAWARE
. This is a preview feature subject to change. See Consecutive days worked for more details.
More reliable constraint matching for shifts across time zones
: Fixed inconsistent tie-breaking for shift-related constraints when shifts spanning different UTC offsets represented the same time instant.

Pick-up and Delivery Routing (v1 | Stable)

Maintenance release
: This release updates dependencies to deliver security updates and minor performance improvements.

Please let us know if you have feedback.