Multi-Horizon Production Planning for High-Complexity Glass Manufacturing
Replace manual, spreadsheet-driven planning with a constraint-aware AI engine that optimizes across every furnace, line, and shift — in real time.
Case Background
- A large glass manufacturer running continuous production across multiple shared furnaces and downstream lines, managing a complex product mix with strict delivery commitments. Production planning was handled manually through spreadsheets — requiring planners to balance demand, capacity constraints, product-line compatibility, and changeover sequences across quarterly, monthly, weekly, and daily horizons simultaneously.
Business Challenge
-
The client faced compounding inefficiencies driven by shared-asset complexity:
- No structured way to translate demand into capacity-feasible plans across multiple horizons.
- Manual changeover sequencing leading to avoidable losses on shared furnaces and lines.
- Limited ability to incorporate real-time execution signals — downtime, scrap, and actual output — into daily replanning.
- No visibility into constraint trade-offs, scenario comparisons, or plan audit history. The result was frequent firefighting, unstable schedules, and missed delivery commitments.
AI-Powered Solution
-
Makkook deployed an integrated Production Planning Optimization Platform — four specialized AI engines unified under a single Planning Cockpit. The platform ingests demand signals, sales orders, inventory, and capacity calendars to generate optimized, feasible plans across all horizons.
The AI Engines in scope include:
1. Demand & Capacity Scenario Engine
2. Multi-Horizon Production Planning Optimizer
3. Changeover-Aware Sequencing Engine
4. Execution-Aware Replanning Engine
The Planning Cockpit provides an interactive Gantt view, constraint explanations, exception alerts, and KPI dashboards — with full plan versioning, scenario comparison, and audit logging across every planning decision.
Business Impact
-
1. Changeover hours reduced through solver-optimized sequencing across shared assets
2. OTIF delivery rate improved through constraint-aware multi-horizon planning
3. Plan stability increased through frozen windows and closed-loop execution feedback
4. Manual planning effort significantly reduced — shifting planners from reactive firefighting to strategic decision-making
5. Full auditability established across all plan versions, constraint overrides, and replanning events.
Stack & Integrations
- Constraint Optimization (MILP/CP-SAT) · Multi-Horizon Planning · Execution-Aware Replanning · Containerized Microservices, integrated with ERP and Order Management, inventory and WIP systems, production reporting, downtime and scrap logs, and furnace and line capacity calendars — deployed on-premise for OT security compliance.