ONE CLOCK · ONE AUTHORITATIVE LEDGER Same work. Two operating systems. Compare manual allocation with Digital Twin control. Follow picking, recovery, delivery and shipping through the same shift.
Digital Twin ↗ INNER Bin study ↗ Priority access study ↗
COUPLED EVENT MODEL synthetic · uncalibrated
View operating systemOld system Digital Twin Both models advance together. The Digital Twin controls dispatch, traffic, recovery and outbound queues.
Scenario controls Existing dispatch · with rules
▶ Play +10 min Run to shift end ↗ Stop calculation ↺ Playback1× 10× 60× 300×
SHARED SIMULATION CLOCK 00:00:00 7.4 h = 7h 24m
Export ledger CSV ↓
Preparing four matched runs…
Retry scenario
Locations confirmed — Counted once, after confirmation Unconfirmed locations — Original unconfirmed quantities stay reserved Original tasks fully picked — Actual task locations; recovered fragments are not extra tasks Picker waiting — Cumulative picker-minutes · traffic + delivery + wrapper queues + STOP
L1 picking · eight sorter delivery sections · rear racks retained Preparing WebGL · schematic available
Picking truck High-bay / waiting Plastic BinDrag to orbit · scroll to zoom
The renderer reads model state; camera movement never advances the clock. Use schematic Retry WebGL
FLOW DIAGNOSTICS Ready to simulate Choose a scenario, then play or advance the same clock in all four runs.
— putaways confirmed
— vehicles in resource queues
Who is waiting?
DELIVERY & CLASSIC OUTBOUND
Delivery lanes and waiting pickers
Keyboard location grid · all 440 addresses in the selected row Arrow keys move by position or level. Home / End selects position 1 / 110. Page Up / Down selects L4 / L1. One Tab stop enters the grid.
AUTHORITATIVE LIVE WORKLOAD Every task has fixed sources, quantities and carriers. 15 voice + 15 scanner. Each location is confirmed only after physical arrival. Delivery remains part of the picker’s cycle.
Same clock
ZoneBoth zones Sorter Classic LinesAll states Unconfirmed Confirmed ← Previous Next →
SORTER / OUTER One OUTER is one carton. Up to 25 locations per task; 1–70 pieces inside each carton, fixed by SKU and pack configuration.
SORTER / INNER Individual pieces, up to 100 locations per Bin task. Bin delivery to the eight sorter rows is a configurable assumption.
CLASSIC / INDIVIDUAL PIECES Up to 100 locations and 200 individual pieces per complete task. Euro / non-Euro pallets go to the wrapper, then direct truck loading.
MATCHED DEADLINE · IDENTICAL SOURCE WORKLOAD Compare decisions through the same shift. Dispatch changes assignment order. Source locations, SKU quantities and carrier IDs remain fixed.
Same clock Rules-off is a deliberate sensitivity scenario. It disables posted-speed and STOP enforcement while retaining shared-resource access. These results do not measure safety, accidents or warehouse productivity.
Delivery, wrapping and truck loading
Current carrier destinations and timestamps Up to 50 active carriers plus recent truck loads. Select an original task for its complete carrier and handover history; CSV includes every record.
INTERRUPTED ORIGINAL TASKS
The unfinished work stays visible.
Worker time conservation Original task ledger ← Previous Next →
Shared event ledger · latest events
Kasynets AI Solutions · operational simulation prototype · · · All inputs synthetic unless explicitly supplied and independently validated.
MODEL SCOPE & EVIDENCE
What this demonstration represents × Operator rules in one coupled cycle The same clock schedules physical arrivals, verification, picking, confirmation, interruptions, carrier collection, delivery queues, placement, wrapping, internal transfer, direct truck loading and shift expiry. Picking complete, delivered, wrapping complete and loaded are separate timestamps. Original task, location, SKU, fixed pack configuration, picker, destination carrier, quantities and confirmation time remain in the ledger.
Layout allocation and usable capacity There are still 100 sorter rows and 50 Classic rows. Exactly eight configurable sorter front sections are delivery areas through the 55/56 split. S001–S008 is the provisional initial assignment. Their rear racks remain. Delivery positions are operational destinations and cannot generate picking stock or tasks. Until measured geometry is supplied, the front delivery allocation replaces all four theoretical rack levels: 66,000 previous theoretical rack addresses become 64,240 usable rack addresses, including 16,060 active L1 pick locations. Other upper slots are inspectable except for released putaways.
Task size and package accounting Sorter OUTER tasks use at most 25 pick locations. One carton contains 1–70 pieces according to a stable synthetic SKU/pack master. Cartons requested/confirmed, pieces per carton and base units are distinct. Bin tasks use up to 100 locations and collect individual pieces. Classic tasks obey both 100 locations and 200 individual pieces for the complete original task; distinct SKU count is independent. The mixed household-goods catalog is illustrative of the described PEPCO operation, not an actual company catalog or WMS feed.
Halfway interruption occurs after floor(actual task locations / 2) confirmations: 25 stops after 12 and leaves 13; 100 stops after 50. Confirmed goods stay on original carriers and unfinished source quantities stay reserved. Recovery returns to the recorded bay and collects the same carriers after handling time. No universal 20-line task is used.
Delivery and empty carrier collection Sorter goods travel to a delivery queue, wait for a finite carrier position, then enter the allocated section and place each identified carrier. Bin delivery to the same eight-row area is an explicit configurable assumption; a separate assumed Bin handover point is available. Lane capacity, placement time and downstream clearance dwell are synthetic inputs. Clearance records a timed transfer to sorter downstream; it does not imply truck loading and never erases confirmed load quantities.
Classic pickers travel to finite wrapper input positions and hand over each pallet. A picker is eligible for its next assignment after the last carrier handover; return travel to the empty-carrier depot and collection still consume its time. One sorter pallet or Bin per task is the initial assumption, configurable to two. Classic retains one Euro and one non-Euro pallet. Empty supply is assumed available. Picker attachment uses half the carrier-collection reference input (15 seconds at the default 30-second reference); high-bay pallet collection uses the full reference. Recovery adds the half-reference attachment to its separate collection handover time.
Wrapping and direct truck loading The wrapper has finite input capacity and a configurable number of machines. Each carrier has feed, wrapping and transfer service. A full loading buffer blocks the wrapper output. Finite loading bays wait for a truck, load each wrapped carrier and record its original ID and base units. Truck arrival time, carrier capacity and replacement delay are configurable assumptions. Wrapping and loading use autonomous handling time; they do not keep the original picker busy after handover. All productive operations stop at the shift deadline, retaining in-flight carriers, input reservations and unfinished outbound work.
Synthetic assumptions and verification limits Counts and task limits follow the operator brief. Dimensions, upper-rack removal in delivery sections, source buffers, handling times, provisional 6/3 km/h speed policy, queue pockets, train clearance, downstream extraction, empty supply, truck availability and provisional LOOADS concepts require measured geometry and approved rules. The model uses exclusive aisle sections and shared road/junction resources, not validated swept-body collision avoidance. Carrier volume/weight, replenishment, batteries, fatigue, failures, pedestrians, actual truck schedules and cross-shift resumption remain outside scope. Passing tests demonstrate consistency, not calibrated productivity or certified safety. Goods meshes are visual cues; ledger quantities are authoritative.
INNER Bin task design case The supplied October aggregate contains 30 PICK and 11 FRAG assignment records, 5,497 individual pieces and 23.8 measured hours. The operator reports 70 source locations per assignment, a 250-piece cap and 20–30 empty cartons requiring cutting. A reconstructed source ledger preserves these totals; actual SKU/order-line data was not supplied. Profile A retains 70-location assignments and cutting; B obeys both 30 locations and 300 pieces per Bin and assigns no cutting to the picker. Order boundaries remain intact by default. Optional compatible-order consolidation assumes shared source/SKU demand, retains every original source-line allocation and is explicitly synthetic. Planning estimates use assumed average travel; applying a profile runs one picker through actual 3D routes, service phases and the shared cutoff. Other pickers remain idle. Carton opening and cutting are separate timed phases and never create picked INNER quantities. These figures illustrate task-design exposure; they do not establish managerial intent.
Digital Twin control and anti-manipulation scope The Old system / Digital Twin mode advances matched source demand and finite resources at one clock. The controller ignores preferred-picker cards; scores released tasks using service, travel, congestion, age, projected category coverage and dense-work exposure; plans source visits; enforces front-aisle directions, rear entry permits, posted speeds and STOP checks; recovers interrupted original carriers; and selects compatible delivery, wrapper and loading queues. It cannot silently remove carton work, create stock or add labour. Front directions and service times are explicit unvalidated assumptions. Complete tasks per paid picker-hour, category output, backlog and shipped goods are shown separately; a gain in every measure is not guaranteed.
Running input rules and original source quantities are pinned. Worker-model commands reject manual priority, false confirmation, wrong-direction and bonus changes. A local SHA-256 audit chain and stock/carrier checks provide tamper evidence. This local browser prototype does not provide production authentication, trusted persistent audit storage or a real WMS/PLC connection. It is a decision-support and integration prototype, not a validated safety or security controller.
Earlier projection model Open the archived sample-projection lab ↗ . It predates the operator’s task-size and delivery requirements. Its average-rate results are not a continuously coupled operational shift and must not be compared as a calibrated baseline.
Measured runtime diagnostics Loading diagnostics…
Frame intervals describe this browser session only. They are not a phone benchmark or a performance guarantee. Rendering and camera state never participate in scheduling or confirmation logic.