Rebulk

Switch Planning · Early access

Every switch, planned in seconds

A live map of every track, spot, and car. The AI plans the work: which cars move, where, in what order, in the fewest moves. Built for industrial yards, plants, shortlines, and switch crews, and set up on your track layout, your spots, and your constraints.

Live yard mapAI move optimizationGenerated switch lists
Loading switch yard scene…

Live simulation: two cars, buried on different tracks, coupled together. The optimizer sequences the trackmobile in the fewest moves; your crew works the same plan as a switch list.

The yard, planned instead of improvised

Switching is a sequencing problem: on fixed track, every move can require three other moves first. Software plans that better than a whiteboard.

Live yard map

Every track, spot, and car in one view: what's occupied, what's ready to work, what's waiting on release. The whiteboard, retired.

AI move optimization

The AI planner weighs every car, track, and constraint, then sequences the fewest moves that satisfy the plan: less engine time, fewer touches per car.

Generated switch lists

Work orders your crew can run: pull this car, spot that one, in an order that doesn't undo itself two moves later.

Spot assignment

Cars land where the work happens: the right bay, the right pit, the right riser, with pinned cars and do-not-work constraints respected.

Dwell and demurrage aware

The planner sees the same per-car clocks as the rest of Rebulk, so cars nearing free-time limits get worked before they start billing.

One record per car

Yard moves write to the same car records as tracking, BOLs, and scale tickets, with no second system to reconcile.

Every unnecessary move is money

A switch move costs engine time, crew time, and fuel, and bad sequencing multiplies it: dig out the wrong car first and you've touched four cars to place one. Cars wait longer for spots, dwell climbs, free time expires, and demurrage starts billing.

Yard supervisors solve this in their heads every morning, until they're off, volumes spike, or the yard fills past what mental math holds. The planner makes the sequencing explicit, repeatable, and one click to re-run when the inbound consist changes.

Tracking cars but not switching them yourselves? Start with Railcar Track & Trace. Both share the same car records.

Frequently asked questions

What is switch planning software?

Switch planning software gives a rail-served facility a live picture of its yard (which cars sit on which tracks and spots) and plans the switching work: which cars move, where, in what order. It replaces the whiteboard, the clipboard switch list, and the supervisor's mental math.

How is switch planning different from a warehouse yard management system (YMS)?

A warehouse YMS manages trucks and trailers in a parking yard. Switch planning deals with railcars on fixed track: cars can't pass each other, one move can require relocating three other cars first, and sequencing determines everything. Rebulk's planner is built for rail: industrial yards, plants, and shortlines, not truck lots.

How does switch move optimization work?

The AI planner treats the yard as a constraint problem: each car's current spot, its destination, track capacities, pinned or do-not-work cars, and crew limits. It searches for the sequence that completes the plan in the fewest moves. Supervisors review and adjust before the crew runs it.

Does Switch Planning work with Rebulk's railcar track & trace?

Yes, they share one record per car. Yard moves update the same data that powers track & trace, BOL and scale ticket automation, dwell clocks, and customer reporting. No integration project.

When is Switch Planning available?

Early access is open now for industrial yards, plants, shortlines, and rail-served facilities that switch their own yards. Request early access or book a demo to see the planner on your track layout.

Bring us your worst switch day.

Early access partners get the planner set up on their own track layout. Bring a recent switch list and we'll show you the sequence the optimizer would have run.