Production Scheduling for Plastic Compounding

If you schedule a plastic compounding plant, you balance purge-driven changeovers across capability-constrained extruders while coordinating mixed batch-and-flow production through multiple stages. This guide shows how to configure Schantt for your compounding operation: model your product classes and their per-class routings, set directional changeover durations, and let the scheduling algorithm sequence jobs to minimise total production time.

This guide follows a fictional composite company built from industry research on plastic compounding; all names, parameters, and figures are illustrative.

Industry context

Plastic compounding takes base polymer resins and transforms them into formulated compounds by mixing in additives, fillers, and reinforcement agents such as glass fibre. The process runs through five sequential stages — premixing (batch), extrusion (flow), pelletizing (flow), classifying (flow), and packaging (flow) — each feeding the next after a material-handling transfer time. What makes compounding scheduling distinctive is the intersection of three factors: asymmetric purge times between colour and polymer families, capability-restricted extruder pools (filled compounds require dedicated lines), and divergent per-class routings where some products skip interior stages entirely.

The composite facility at the centre of this guide processes three product classes — PP-Natural-Unfilled, PA-Black-GF30, and PE-Natural-Unfilled — across thirteen machines arranged in five production stages. PP-Natural-Unfilled runs through all five stages. PA-Black-GF30 is a glass-fibre-reinforced, coloured compound that requires a dedicated extruder line and skips the classifying stage. PE-Natural-Unfilled enters production at the extrusion stage only, bypassing premixing entirely through direct loss-in-weight feed, and also skips classifying. The product mix is approximately fifty percent custom-order, thirty percent make-to-stock standard compounds, and twenty percent toll-compounding work for third-party customers, giving a scheduling workload of varied job sizes, changeover profiles, and routings across a typical week.

Novoplast Compounds runs approximately 85 people at an 8,000 m² facility, making three product classes across five production stages, scheduled by a three-person planning team. Extrusion operates on a continuous 120-hour week — Monday 06:00 through Saturday 06:00 — while packaging runs a two-shift schedule totalling 88 hours per week. Two weekly six-hour screw pulls are staggered across different extruders (one on a Tuesday, one on a Thursday), and a seven-day annual shutdown closes the plant in the last week of December. The planning team manages between six and nine jobs per week, each defined by its product class and quantity, and spends a significant portion of each schedule run manually sequencing jobs to avoid time-consuming purge events.

Process overview

flowchart LR
    PM["Premixing<br/>(Batch)"] --> EX["Extrusion<br/>(Flow)"]
    EX --> PL["Pelletizing<br/>(Flow)"]
    PL --> CL["Classifying<br/>(Flow)"]
    PL --> PK["Packaging<br/>(Flow)"]
    CL --> PK

Five-stage production flow from premixing through packaging. Pelletizing has a direct bridge to packaging for product classes that skip classifying.

PE-Natural-Unfilled enters at Extrusion (premixing skipped via direct loss-in-weight feed). PA-Black-GF30 and PE-Natural-Unfilled skip Classifying — their pellets exit the pelletizer sufficiently clean. Bridging transfer times connect across the skipped spans.

Scheduling challenges and how Schantt handles them

This guide assumes the primary demand signal is customer orders — jobs with a product, quantity, and due window that arrive as the schedule horizon opens. If your facility schedules from make-to-stock targets or toll-compounding contracts, the same configuration applies; the demand shape differs but the modelling steps are the same.

Schantt's scheduling algorithm minimises total production time for the set of jobs scheduled, working forward from a start date you choose. For this guide, the practical horizon is a rolling week — enough to demonstrate the changeover dynamics and calendar interaction that make plastic compounding scheduling distinctive.

Schantt provides two optimisation modes. Auto mode explores both job sequence and machine assignment simultaneously to find the schedule with the shortest total production time. Semi-Auto mode optimises machine assignments while keeping the job sequence you specify.

What Schantt handles well

• Sequential multi-stage production — compound products run through an ordered pipeline of premixing (batch), extrusion (flow), pelletizing (flow), classifying (flow), and packaging (flow), each stage feeding the next after a transfer time.
• Multi-machine stages with capability-restricted assignment — not every extruder can run every compound; you enter rates only for compatible product classes, and Schantt assigns jobs to eligible machines automatically.
• Mixed batch-and-flow pipelines — premixing is batch (fixed cycle per load) while extrusion, pelletizing, classifying, and packaging run at continuous rates — Schantt handles both types in a single route.
• Multi-product routing with stage skipping — some product classes skip premixing (direct loss-in-weight feed) or classifying — per-class routing omits the skipped stages and bridges the gap with transfer times.
• Sequence-dependent changeovers (directional purge durations) — colour and polymer-family transitions have asymmetric purge times; directional changeover settings let the scheduling algorithm cluster similar jobs to minimise total changeover time.
• Shift-aware availability and scheduled downtimes — twenty-four-hour continuous extrusion with weekend screw pulls is modelled through calendars and machine-downtime windows, so the schedule respects actual working time.

How Schantt handles each challenge

1. Directional changeover durations for asymmetric purge times.

• In a compounding plant with three product classes, the changeover time between two classes on a given machine depends on which class is leaving and which is arriving. On the general-purpose extruders, for example, a transition from PP-Natural-Unfilled to PE-Natural-Unfilled takes 60 minutes of purge time, while the reverse direction takes 45 minutes — a 15-minute difference that repeats across every job boundary in the week. A planner scheduling manually may not consistently favour the shorter direction, especially as the job queue changes.
• Schantt models changeovers as per-pair directional durations entered on the machine's detail page. Each (from-class, to-class) pair on a machine gets its own value — 60 minutes one way, 45 minutes the other — and the scheduling algorithm accounts for the actual transition time at every job boundary. In Auto mode the algorithm explores job sequences and favours those with shorter total changeover time, naturally clustering similar product classes and avoiding directional pairs with the longest purge times. The changeover appears as its own labelled segment on the Gantt, so the planner can see where each purge window falls.

2. Capability-restricted machine assignment for segregated extrusion.

• Machine eligibility in compounding is determined by the physical segregation of production lines. Extruder D runs only PA-Black-GF30 because it handles filled and coloured material that would contaminate the unfilled product stream. The general-purpose line — Extruders A, B, and C — runs only unfilled PP and PE. No machine in this plant can process both families of compounds without an extended purge that would make such sharing impractical. Pelletizer 2 can handle PA material alongside PP and PE, so pooling is permitted there; the restriction is specific to extrusion.
• Schantt enforces machine eligibility through the simple principle that only rates you enter are considered. For Extruders A, B, and C you enter throughput rates for PP-Natural-Unfilled and PE-Natural-Unfilled only. For Extruder D you enter throughput for PA-Black-GF30 only. The scheduling algorithm never assigns a job to a machine that has no rate entered for that job's product class, so segregation is inherent in the data — no separate rule or constraint is needed. The same mechanism applies at every stage: downstream machines with restricted eligibility work the same way.

3. Mixed batch-and-flow pipeline with per-class stage skipping.

• The three product classes at this facility follow different paths. PP-Natural-Unfilled runs all five stages: batch premix feeds the extruder hopper, then the melt strand goes through pelletizing, classifying, and bagging. PA-Black-GF30 also starts with batch premix but skips classifying — its pellets are clean enough after pelletizing to go directly to packaging. PE-Natural-Unfilled skips premixing entirely (the polymer powder feeds directly into the extruder by loss-in-weight system) and also skips classifying. This means the PE class has only three stages in its routing, yet it shares extruder and packaging capacity with PP and PA jobs that have longer routing chains.
• Schantt handles mixed batch-and-flow production by letting each stage be typed independently — Premixing is a batch stage, while Extrusion, Pelletizing, Classifying, and Packaging are flow stages — and each product class defines its own routing through them. The three classes use three different routing lengths. Transfer times bridge the gaps where stages are skipped: a 20-minute pelletizing-to-packaging transfer connects across the classifying skip. At the premix-to-extrusion handoff, the per-class partial-transfer setting allows the extruder to begin processing the first 300 kg of the batch while the mixer continues the next load, without waiting for the full premix quantity to finish.

4. Shift-aware availability with disparate calendars and planned maintenance.

• The calendar mismatch between upstream and downstream stages is the most common hidden constraint in compounding scheduling. A planner building a schedule manually can sequence extrusion jobs efficiently only to discover that packaging cannot start until Monday because the bagging lines do not run on a Saturday night shift. Similarly, if the planner does not account for Extruder A being offline for six hours on Tuesday afternoon for its screw pull, the schedule will assign a job that cannot physically run.
• Schantt supports multiple calendars within a single schedule. Extrusion, pelletizing, and classifying use the Continuous Production calendar (120 hours per week). The three packaging machines use the Packaging Shift calendar (88 hours per week). The scheduler computes available working time per calendar, so a packaging job that spans Saturday evening clamps forward to Monday morning automatically. Screw pulls are entered as machine-downtime events on their respective days, and the annual shutdown is entered as a factory-wide downtime window. The Gantt renders shaded overlays for both non-working calendar gaps and downtime events, so the planner can see at a glance why extrusion production pauses are scheduled around Tuesday's maintenance and why packaging jobs never appear after Saturday morning shift end.

What to model in Schantt

These are the five first-class entities you create to represent your compounding plant in Schantt.

Step-by-step setup

1. Create the five stages and set transfer times. Add Premixing (batch), Extrusion (flow), Pelletizing (flow), Classifying (flow), and Packaging (flow) in process order. On each stage's detail page, enter the transfer time to the downstream stage:

• Premixing → Extrusion: 15 minutes
• Extrusion → Pelletizing: 10 minutes
• Pelletizing → Classifying: 15 minutes
• Pelletizing → Packaging: 20 minutes (bridge route for classes that skip classifying)
• Classifying → Packaging: 10 minutes

2. Add the machines to each stage. Assign all thirteen machines to their respective stages. Each machine inherits its stage's default calendar initially; the calendar override is configured in step 6.

• Premixing: Mixer A, Mixer B
• Extrusion: Extruder A, Extruder B, Extruder C, Extruder D
• Pelletizing: Pelletizer 1, Pelletizer 2
• Classifying: Classifier 1, Classifier 2
• Packaging: Bagging Line 1, Bagging Line 2, Bulk Load-Out

3. Create the three product classes and define per-class routing. Add PP-Natural-Unfilled, PA-Black-GF30, and PE-Natural-Unfilled. On each class's detail page, define its routing through the stages. Enable partial transfer on the premix leg for the classes that use it:

• PP-Natural-Unfilled: Premixing (partial transfer enabled, 300 kg minimum) → Extrusion → Pelletizing → Classifying → Packaging
• PA-Black-GF30: Premixing (partial transfer enabled, 300 kg minimum) → Extrusion → Pelletizing → Packaging (skip Classifying)
• PE-Natural-Unfilled: Extrusion → Pelletizing → Packaging (skip Premixing and Classifying)

4. Add one representative product per class. Create a single product under each product class — three products total. These are the reference items that carry the class routing onto each job.

5. Set machine capacity parameters and changeovers. On each machine's detail page, enter throughput rates for flow stages and batch parameters for batch stages. This step requires the product classes from step 3 to be created first, because throughput and changeover entries reference them.

• Premixing mixers (Mixer A, Mixer B): batch cycle 10-12 minutes, batch size 300 kg — for PP and PA classes only (PE-Natural-Unfilled skips this stage)
• Extruders A, B, C (general-purpose): PP-Natural-Unfilled at 500-550 kg/h, PE-Natural-Unfilled at 600-650 kg/h — no rate entered for PA class
• Extruder D (filled-dedicated): PA-Black-GF30 at 400 kg/h only
• Pelletizer 1, Pelletizer 2: PP and PE at 1,500 kg/h; Pelletizer 2 also handles PA at 1,200 kg/h
• Classifier 1, Classifier 2: 1,000 kg/h for PP-Natural-Unfilled only
• Packaging stations: Bagging lines at 800 kg/h, Bulk Load-Out at 1,500 kg/h — all three classes at these rates
• Directional changeovers on premixing: on Mixer A and Mixer B, enter PP→PA at 25 minutes and PA→PP at 20 minutes as separate entries (a brief vessel clean between unfilled and glass-fibre-reinforced material)
• Directional changeovers on extrusion: on Extruders A, B, and C, enter PP→PE at 60 minutes and PE→PP at 45 minutes as separate entries. Pelletizer 2 and packaging stations also have per-pair changeover entries between the classes they share

6. Configure calendars, exceptions, and downtimes. Set Continuous Production as the default calendar. Assign the Packaging Shift calendar to Bagging Line 1, Bagging Line 2, and Bulk Load-Out. Add the three calendar exceptions and three downtime windows:

• Continuous Production: Monday 06:00 to Saturday 06:00 (120 hours per week) — applies to premixing, extrusion, pelletizing, and classifying
• Packaging Shift: Monday to Friday 06:00-22:00, Saturday 06:00-14:00 (88 hours per week) — applies to packaging machines
• Exceptions: New Year's Day (non-working), International Workers' Day (non-working), Christmas / year-end break (non-working)
• Downtimes: Extruder A screw pull (6 hours on a Tuesday), Extruder D screw pull (6 hours on a Thursday, staggered), annual shutdown (7 days, December 24-31, all machines)

For step-by-step instructions on configuring each of these in Schantt, see the Schantt documentation.

Common mistakes

1. Using a single blanket changeover time instead of per-pair directional durations. When all transitions are entered as the same value on every extruder, the scheduling algorithm treats a PP-to-PE move as equivalent to a PE-to-PP move, even though the actual purge times differ by 15 minutes or more. The optimisation benefit of clustering similar runs is lost. Fix: Enter each direction as a separate changeover entry on the machine, using the measured purge duration for each (from-class, to-class) pair.

2. Placing incompatible product classes into the same machine pool. Assigning a throughput rate for PA-Black-GF30 on Extruder A alongside PP and PE implies that Extruder A can run all three classes. The scheduling algorithm will assign PA jobs to Extruder A when it finds a capacity advantage, creating a schedule that cannot be executed on the floor without an extended (120+ minute) purge — or may contaminate the product. Fix: Enter throughput rates for filled compounds only on the dedicated extruder and leave them off the general-purpose lines. The algorithm will never assign a job to a machine that has no rate for that class.

3. Using a single calendar across all stages. When every machine shares the same twenty-four-hour calendar, the schedule assumes packaging is available through Saturday night and produces a plan that looks feasible but stops on the packaging floor when the bagging lines go off shift. Fix: Assign the Packaging Shift calendar to the three packaging machines. The schedule then computes packaging jobs using only the actual 88-hour working week and surfaces the weekend gap on the Gantt.

4. Modelling stage skips as zero-time empty stages. Adding a Premixing stage with zero duration to PE-Natural-Unfilled's routing produces an unnecessary operation on the Gantt and may introduce unintended delays as the forward-only transfer chain adds handling time for a stage the product never physically visits. Fix: Omit the skipped stages from the class's routing definition. The bridging transfer time — pelletizing to packaging at 20 minutes — connects the routing automatically across the gap.

5. Setting symmetrical changeover durations despite asymmetric purge reality. Using the same duration in both directions hides the time penalty of unfavourable job sequences. A scheduling algorithm that sees 45 minutes in both directions cannot distinguish between a favourable PE-to-PP transition (45 minutes) and an unfavourable PP-to-PE transition (which actually takes 60 minutes), so it has no information to favour one sequence over the other. Fix: Measure and enter the actual purge time for each direction. The asymmetry becomes data the algorithm can use to improve the schedule.

What a good schedule looks like

A well-configured Schantt schedule for this compounding plant replaces manual sequencing with algorithm-optimised job ordering and machine assignments, turning hidden calendar gaps and changeover asymmetry into visible, actionable constraints.

Before (manual scheduling):
- 4 to 8 hours or more of non-productive purge time per week from unfavourable job sequences — the planner may cluster similar classes but has limited capacity to explore alternative orderings across 6 to 9 weekly jobs
- Extrusion line utilisation estimated at 60 to 75 percent of available capacity, with idle gaps between jobs that a manual Gantt does not help fill
- Packaging bottlenecks discovered only after the schedule is built — extrusion jobs advance continuously while packaging stands idle on Sunday, but the misalignment is not visible during planning
- Production completion on the schedule may differ from readiness to ship by 24 to 48 hours for QC testing; this release step is managed outside the scheduling system

After (Schantt Auto mode):
- Changeover time reduced by approximately 30 to 50 percent as the algorithm clusters similar product classes and avoids long directional transitions — the 60-minute PP-to-PE purge on the general-purpose extruders occurs less frequently
- Extrusion utilisation improved to 75 to 85 percent or higher as the algorithm reduces idle gaps between jobs
- Calendar-aware scheduling ensures packaging jobs land only during the 88-hour packaging shift — the Gantt shows shaded overlays where the packaging calendar is closed, making the mismatch visible before the schedule is finalised
- Staggered screw pulls are automatically respected: jobs on Extruder A pause for its Tuesday six-hour maintenance window and resume on Wednesday; no manual re-routing is needed