Constraint-Based ATP

The core algorithmic component that processes a network of hard and soft constraints to find a feasible promise date. Unlike sequential rule-based checks, the solver evaluates all limitations simultaneously, backtracking when conflicts arise to ensure the final commitment is achievable across all resources.

Simultaneously evaluates three critical constraint dimensions:

• Material Constraints: On-hand inventory, scheduled receipts, and BOM component availability
• Capacity Constraints: Finite machine hours, labor shifts, and tooling availability at each work center
• Transportation Constraints: Carrier capacity, transit times, and shipping calendars The solver only commits when all three dimensions align.

The solver employs two strategic approaches to find a valid slot:

• Backward Scheduling: Starts from the requested delivery date and works backward, reserving capacity and material at the last possible moment to minimize inventory holding costs
• Forward Scheduling: When backward scheduling fails due to a constraint violation, the solver shifts forward to find the earliest feasible date after the request The system automatically selects the optimal strategy based on the constraint profile.

Enables planners to test hypothetical scenarios without affecting live commitments. Common simulations include:

• Adding a new high-priority order to a fully loaded production line
• Simulating a supplier delay on a critical component
• Evaluating the impact of a machine breakdown on existing order promises Results show which orders would be affected and provide alternative fulfillment options.

When no feasible solution exists, the solver uses constraint relaxation to identify the binding constraint preventing fulfillment. It then propagates the impact across dependent orders, showing planners exactly which bottleneck—a specific work center, a late supplier shipment, or a transportation lane—must be resolved to meet the commitment date.

Constraint-Based ATP is tightly coupled with the finite capacity scheduling engine. When the solver reserves capacity for a new order, it updates the production schedule in real-time, ensuring subsequent ATP checks see the newly constrained environment. This closed-loop integration prevents overcommitment and maintains schedule fidelity across all order promising operations.

A predefined policy that dictates the sequence of supply locations the constraint solver evaluates. For a given customer region, the rule might specify:

• Primary: Regional DC in Chicago
• Secondary: Plant in Monterrey
• Tertiary: Any global DC

The constraint solver applies these rules while simultaneously checking capacity and material constraints at each node.

The algorithmic extension where the constraint solver evaluates all possible combinations of supply sources to fulfill a multi-line order. Rather than promising each line independently, the solver finds the global optimum that minimizes total landed cost or maximizes margin, even if it means splitting a single order across three different shipping points.

The financial layer that sits atop Constraint-Based ATP. Once the solver identifies a feasible fulfillment plan, PTP logic evaluates its cost-to-serve—including freight, handling, and potential expediting charges—against the order's margin. If the cost exceeds profitability thresholds, the solver may reject the promise or propose an alternative, higher-margin fulfillment path.

The foundational calculation that feeds the constraint solver. Netting subtracts gross demand (sales orders, forecasts, safety stock) from scheduled receipts (purchase orders, production runs, in-transit inventory) to compute the projected available balance at each time bucket. The constraint solver uses this time-phased inventory projection as the material constraint input.