Deterministic planning assumes a perfect world. The shopfloor is anything but. This paper introduces a novel Simulation-Optimization approach that merges Discrete Event Simulation (DES) with Constraint Programming (CP) to dynamically execute "what-if" scenarios, eradicate bottlenecks, and achieve true Just-in-Time production.
Most modern ERP and APS (Advanced Planning and Scheduling) systems rely on deterministic models. They generate a daily schedule assuming standard processing times, zero machine breakdowns, and perfect raw material availability. The moment a tool breaks or a setup takes 15 minutes longer than anticipated, the static schedule shatters. This stochastic noise leads to massive Work-in-Progress (WIP) buffers, delayed downstream logistics, and missed SLA targets.
To optimize under uncertainty, Presolve Labs deploys a hybrid architecture. We pair a high-fidelity Discrete Event Simulation (DES)—acting as a digital twin of the shopfloor—with a lightning-fast Constraint Programming (CP) solver.
The visual dashboard to the right demonstrates this shift. Under standard heuristic routing, workpieces blindly follow a fixed path to Machine 3, causing a cascading queue that blocks upstream operations.
By deploying the DES-CP hybrid, the system achieves true Just-In-Time (JIT) flow. Inventory holding costs plummet as WIP is reduced by over 40%. Because parts reliably exit the shopfloor exactly when needed, follow-up logistics can operate with tighter margins, raising On-Time In-Full (OTIF) delivery metrics to 99%+.
Fig 1. Simulated material routing. Contrast the rigid scheduling (leading to M3 bottleneck) against CP-optimized dynamic routing (balancing M3 and M4) for optimal throughput.