Where warehouse decisions go wrong without simulation
Three decisions in warehouse and logistics consistently produce poor outcomes when made without simulation: AGV fleet sizing, throughput capacity planning, and daily shipping scheduling.
AGV fleet sizing made from formulas or vendor quotes consistently misses traffic interaction effects — the non-linear congestion that makes a 20-vehicle fleet perform worse than a 16-vehicle fleet in certain layouts. Throughput capacity planning that ignores stochastic failure and peak demand behaviour produces a facility that works in the spreadsheet and fails on the busiest Monday of Q4. Daily scheduling that treats the plan as fixed ignores every overnight event that made yesterday's plan obsolete.
The common thread: warehouse systems are non-linear. Small changes in layout, fleet size, or scheduling rules produce disproportionate effects on throughput. The only way to predict those effects correctly is to model the interactions — which is what simulation does.
Three simulation applications for warehouse & logistics
AGV Fleet Sizing
Discrete-event simulation of your full vehicle fleet — traffic, charging, ASRS interaction, and peak demand. Determines the minimum fleet that meets your throughput target with a defined confidence level.
Intralogistics Throughput
Sorter capacity, conveyor buffer sizing, dock scheduling, and pick zone layout — modelled as a complete system. Identify where throughput is lost and what changes produce the highest return.
Planning Digital Twin
Live connection to your WMS or ERP. Every morning the model reads current order queue, staffing, and dock availability — and outputs a revised shipping schedule with on-time probability per order.
AGV fleet sizing — why formulas fail
The instinct is to calculate: trips per shift ÷ trips per vehicle = fleet size, plus a buffer. This calculation is wrong because it assumes vehicles work independently. In a real warehouse floor, AGVs share corridors and compete for charging stations. A fleet sized to theoretical throughput may deliver 65% of that throughput in practice due to conflicts — or deliver 110% because the layout happens to have ideal separation between charging and pick zones.
We build AGV simulations in Simio using the SimulateFirst AGV Framework, which models battery management, traffic management, ASRS interfaces, and dispatch logic as explicit agent behaviours. A typical study runs 5–8 fleet size and configuration scenarios and delivers a throughput curve showing exactly where additional vehicles stop adding capacity.
Read more: How many AGVs does your warehouse actually need?
Intralogistics throughput — modelling the full system
Distribution centres fail at the interfaces between subsystems: sorters that outrun replenishment, conveyor buffers that block when three zones hit peak simultaneously, dock doors that become the binding constraint at 14:00 every day when the outbound shift begins mid-inbound window.
We model these interactions explicitly in Simio or AnyLogic, capturing:
- Sorter throughput with realistic recirculation rates and induction timing
- Pick zone throughput by SKU profile, pick density, and wave structure
- Conveyor buffer capacity and the propagation of starvation upstream
- Dock scheduling: inbound vs. outbound conflict, truck arrival variability
- Staff allocation across zones under variable demand
Production planning digital twin — daily schedule that reflects reality
Static warehouse scheduling assumes the plan from yesterday still holds. It doesn't. Overnight orders changed. A picker called in sick. The inbound truck with the priority SKU was delayed. A scheduling system that can't see any of this will generate a Gantt chart that bears no resemblance to what's actually achievable today.
The SimulateFirst planning digital twin connects to your WMS or ERP and reads current state every morning. It runs stochastic simulation of the next 2–5 days and outputs a revised shipping schedule with delivery probability per order — so you know which orders to protect, which to flag to customers, and where to reallocate labour.
Typical integration: REST API or direct database connection to WMS; output via API back to planning system or dashboard.