Production Scheduling for Agrochemicals

Production scheduling for agrochemicals involves planning multi-formulation production across shared batch-and-flow equipment where product classes follow divergent routes and changeover times depend on the transition direction. This guide shows how to model those constraints in a scheduling platform and configure a realistic scenario step by step.

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

Industry context

Agrochemical formulation transforms active ingredients and inert carriers into stable, application-ready products through processes that differ fundamentally by product class. Suspension concentrates (SC) require wet milling to reduce particle size; emulsifiable concentrates (EC) dissolve active ingredients in solvents without milling; water-dispersible granules (WG) need extrusion and drying. Each class imposes distinct constraints on shared equipment — from contamination-driven cleaning between incompatible classes to the seasonal demand surge that follows the planting cycle.

The composite facility that anchors this guide, Meridian Crop Protection, operates 85 people at a 4,500-square-metre plant that produces three product classes — suspension concentrate, emulsifiable concentrate, and water-dispersible granule — across six production stages, coordinated by a planning team of three. Annual volume runs at approximately 3,500 tonnes of SC, 3,000 tonnes of EC, and 1,500 tonnes of WG. Pre-mix cycles range from 30 to 90 minutes for SC and 45 to 75 minutes for WG; wet mills process at 200 litres per hour for SC and 150 litres per hour for WG. Compounding spans 30 to 120 minutes for SC and 30 to 60 minutes for EC. Liquid fill lines run at 45 containers per minute for small packs (0.25 to 2 litres) and 12 containers per minute for large packs (5 to 20 litres); the dry fill line handles 1 to 25 kilogram bags at 10 containers per minute. Quality control holds apply a minimum of 4 hours for SC and EC and 6 hours for WG. The plant runs a standard 40-hour week during off-peak periods and switches to a 96-hour week (double shift, six days) from mid-January to mid-April to meet seasonal demand.

Process overview

flowchart LR
    PM["Pre-mix / Dispersion"] --> WM["Wet Milling"]
    WM --> CP["Compounding"]
    CP --> LF["Liquid Filling"]
    WM --> GD["Granulation & Drying"]
    GD --> DF["Dry Filling"]

Six production stages at the composite facility. SC and WG share pre-mix and wet milling then diverge; EC enters at compounding, skipping both upstream stages.

Note: EC skips Pre-mix / Dispersion and Wet Milling entirely, entering the process at Compounding. WG routes through Granulation & Drying and Dry Filling, bypassing Compounding and Liquid Filling.

Scheduling challenges and how Schantt handles them

Agrochemical production scheduling at this facility is driven by seasonal demand — the schedule must accommodate a pronounced peak from mid-January to mid-April when the 40-hour standard week shifts to a 96-hour double-shift calendar. Readers whose primary constraint differs, such as raw-material availability or regulatory campaign windows, should note that the scheduling approach described here still applies; the driver simply shifts to whichever constraint governs their operation. The scheduling algorithm minimises total production time, working forward from a specified start date, and this guide assumes a practical horizon of four to eight weeks. Auto mode lets the algorithm decide job sequence and machine assignments together; Semi-Auto mode preserves a fixed production order while optimising machine assignments within it.

What Schantt handles well

• Ordered multi-stage routing with stage skipping — each formulation class (SC, EC, WG) follows exactly the stages it needs; classes that skip stages or diverge after shared upstream stages each trace their own path through shared equipment.

• Multi-stage coordination with transfer delays — handoff timing between pre-mix, wet milling, compounding, and filling is chained correctly; fast stages wait for slower ones; material transfers carry configurable delays including quality-control hold.

• Parallel-machine assignment — with multiple vessels, mills, and filling lines per stage, the system assigns jobs across capable machines to minimise total production time.

• Mixed batch-and-flow pipelines — batch stages (pre-mix, compounding) and flow stages (wet milling, liquid filling) run in the same route; partial transfers let downstream stages begin on the first fraction while upstream processing continues.

• Directional, sequence-dependent changeover matrices — asymmetric cleaning times by from-product-class to to-product-class pair are modelled per machine; the algorithm favours sequences that reduce total changeover time.

• Shift-aware calendars with seasonal variation — separate calendars for single-shift off-season and double-shift pre-season are assigned by date range; per-machine calendar overrides and scheduled downtimes subtract maintenance windows.

How Schantt handles each challenge

1. Sequence-dependent changeovers across incompatible classes.
- Changeovers between SC and WG on the pre-mix dispersers and wet mills consume 90 to 180 minutes per transition, while within-class changeovers take only 15 to 45 minutes. Cross-class cleaning between SC and EC on compounding kettles adds 90 to 120 minutes per direction. These directional differences mean that the sequence in which jobs run directly determines how much total time is lost to cleaning — changeover-related delays can account for 15 to 25 percent of available production time.
- The planner enters changeover durations per machine for each from-to pair, informed by the facility's contamination-risk and good manufacturing practice policy. The scheduling algorithm then sequences jobs to favour shorter-changeover transitions across the shared machine sets — for example, running two SC batches back to back before switching to WG — reducing the total time spent on cleaning without the planner having to sequence every transition manually. Schantt schedules per-machine changeovers independently; the planner reviews shared-resource cleaning windows such as a common CIP skid on the Gantt to confirm no unintended overlap.

2. Quality-control hold extending across non-working days.
- Finished batches in compounding must wait through a QC hold — 4 hours minimum for SC and EC, 6 hours for WG — before material is available for filling. When a batch finishes late on a Friday, that hold stretches across the weekend: a compounding batch completing at 15:00 on Friday does not clear its hold until 09:00 on Monday, an effective extension of roughly 66 hours instead of the nominal hold duration. Without calendar-aware timing, the planner must pad manually or risk filling lines sitting idle.
- The QC hold duration is configured as a transfer time from compounding to liquid filling (SC, EC) or from granulation and drying to dry filling (WG) on each stage's detail page. Because transfer times are calendar-aware, the hold window extends automatically across non-working days — the schedule shows the material as available for filling only after the full calendar-elapsed hold elapses. The actual release decision remains a manual quality process outside the scheduling model.

3. Wet mill bottleneck with divergent downstream routes.
- The two horizontal bead mills serve both SC and WG products. SC milling runs at 200 litres per hour per mill; WG milling is slower at 150 litres per hour because the higher solids loading demands more residence time. A delay of 2 hours on the mills can ripple into a 4 to 6 hour total extension by the time downstream compounding, granulation, or filling stages are affected. The bottleneck is compounded by the fact that SC and WG diverge after milling — SC moves to compounding then liquid filling, while WG moves to granulation and drying then dry filling — so a mill queue that serves both classes simultaneously forces trade-offs about which downstream path to feed first.
- With two bead mills in a flow stage, the algorithm can assign each milling job to whichever mill is available, balancing the load between the two machines. Partial transfers from pre-mix to wet milling and from wet milling to compounding (set at 1,000 kilogram increments for SC and WG) let downstream compounding or granulation begin once the first partial batch clears the mill, rather than waiting for the full milling quantity to finish. The planner sets conservative throughput values per class to account for grindability variation and revisits these values when active ingredient source or crystal morphology changes.

4. Seasonal demand requiring a calendar switch.
- The plant operates a standard 40-hour week (single shift, Monday to Friday) for most of the year but needs to ramp to a 96-hour week (double shift, Monday to Saturday) from mid-January to mid-April to meet the pre-planting application window. Manually adjusting machine availability across this transition is error-prone — the planner must remember which machines are affected, ensure the shift pattern applies to the correct date range, and revert when the peak ends. A missed transition can delay orders or leave scheduled operations in non-working time.
- The planner creates two calendars — Standard (40 hours per week) and Peak Season (96 hours per week) — and assigns the Peak calendar to the mid-January to mid-April date range through a schedule calendar period. Calendar exceptions for New Year's Day and International Workers' Day mark those dates as non-working across all calendars. A year-end plant shutdown and a planned summer maintenance window on one bead mill are entered as machine downtimes, subtracted automatically from available capacity. The capacity adjustment itself is a business decision informed by demand forecasting done outside the scheduling platform.

What to model in Schantt

The following table lists the first-class entities needed to represent the Meridian Crop Protection scenario in the scheduling platform.

Step-by-step setup

1. Create the stages and set transfer times. Create the six stages in positional order — Pre-mix / Dispersion (batch), Wet Milling (flow), Compounding (batch), Granulation & Drying (batch), Liquid Filling (flow), Dry Filling (flow). On each stage's detail page, configure the transfer times between successive stages:

• Pre-mix to Wet Milling — 15 minutes (physical slurry transfer)
• Wet Milling to Compounding — 15 minutes (milled slurry to kettle)
• Wet Milling to Granulation & Drying — 15 minutes (WG paste to granulator)
• Compounding to Liquid Filling — 240 minutes (QC hold for SC and EC; calendar-aware, extends across non-working days)
• Granulation & Drying to Dry Filling — 360 minutes (QC hold for WG; calendar-aware)

The Compounding to Liquid Filling transfer bridges the skipped stages for EC, which enters the route at compounding. Because the pre-mix and wet milling stages exist in the stage list but EC never visits them, no phantom operation rows are created on the schedule.

2. Add the machines to each stage. Assign the eleven machines to their respective stages:

• Pre-mix: High-Speed Disperser 1, High-Speed Disperser 2
• Wet Milling: Horizontal Bead Mill 1, Horizontal Bead Mill 2
• Compounding: Compounding Kettle 1, Compounding Kettle 2
• Granulation & Drying: Extruder Granulator, Fluid-Bed Dryer
• Liquid Filling: Liquid Fill Line A, Liquid Fill Line B
• Dry Filling: Dry Fill Line

This gives each stage between one and two parallel machines; the granulation-and-drying stage has two machines that run sequentially per batch (extrusion, then drying), modelled as a single batch stage where the effective cycle duration is the longer of the two per-batch cycle times.

3. Create the product classes and define per-class routing. Create three product classes: Suspension Concentrate (litre units), Emulsifiable Concentrate (litre units), and Water-Dispersible Granule (kilogram units). On the detail page for each class, set its routing — the ordered list of stages it actually visits:

• SC (4 stages): Pre-mix / Dispersion → Wet Milling → Compounding → Liquid Filling. Enable partial transfers on the Pre-mix to Wet Milling leg and the Wet Milling to Compounding leg, each at 1,000 kilogram increments. The compounding and liquid filling legs do not use partial transfers.
• EC (2 stages): Compounding → Liquid Filling. No partial transfers — dissolution is complete before the batch moves.
• WG (4 stages): Pre-mix / Dispersion → Wet Milling → Granulation & Drying → Dry Filling. Enable partial transfers on the Pre-mix to Wet Milling leg at 1,000 kilogram increments. The remaining legs do not use partial transfers.

4. Add one representative product per class. Create three products, each belonging to one class: Azoxystrobin 250 SC, Chlorpyrifos 480 EC, and Atrazine 90 WG. Each inherits its routing and machine configuration from its product class. Assign a distinct display colour per product for Gantt readability.

5. Configure machine capacity parameters and changeovers. On each machine's detail page, set the batch or flow parameters for the product classes the machine handles, then define the directional changeover times per machine:

Pre-mix (dispersers — batch): Cycle duration 60 minutes, batch size 3,000 kg for SC; cycle duration 75 minutes, batch size 3,000 kg for WG. Changeovers: SC to WG — 135 minutes; WG to SC — 90 minutes.

Wet Milling (bead mills — flow): Throughput 200 litres per hour for SC; 150 litres per hour for WG. Changeovers: SC to WG — 135 minutes; WG to SC — 90 minutes.

Compounding (kettles — batch): Cycle duration 60 minutes, batch size 3,000 kg for SC; cycle duration 45 minutes, batch size 3,000 kg for EC. Changeovers: SC to EC — 120 minutes; EC to SC — 90 minutes.

Granulation & Drying (batch): Cycle duration 360 minutes on the extruder granulator per 3,000 kg batch, followed by 240 minutes on the fluid-bed dryer per 3,000 kg batch. The stage duration for one full batch is the long-path machine's time (360 minutes granulation, then 240 minutes drying — these run sequentially, not in parallel, as one batch-stage operation). WG is the only class routed to this stage; no changeover entries needed.

Liquid Filling (flow): Liquid Fill Line A at 675 litres per hour for SC and EC; Liquid Fill Line B at 3,600 litres per hour for SC and EC. Changeovers: SC to EC — 45 minutes; EC to SC — 45 minutes (both fill lines). The dry-fill machine is set up to handle only WG; liquid-fill machines handle only SC and EC — this machine eligibility is implicit in which throughput entries exist for each machine.

Dry Filling (flow): Dry Fill Line at 3,000 kilograms per hour for WG only. No changeover entries needed (single class on a single machine).

6. Configure calendars, exceptions, and downtimes. Create the Standard calendar (40 hours per week: Monday to Friday, 08:00 to 17:00 with a one-hour lunch) and mark it as the default. Create the Peak Season calendar (96 hours per week: Monday to Saturday, 06:00 to 22:00) and assign it to the mid-January to mid-April date range in the schedule using a schedule calendar period. Add calendar exceptions for New Year's Day (1 January) and International Workers' Day (1 May), both non-working. Add two machine downtimes: a year-end plant shutdown from 24 December 18:00 to 1 January 06:00 affecting all machines, and a planned summer maintenance window on Horizontal Bead Mill 1 from 14 July 06:00 to 15 July 22:00 for bead change and seal replacement.

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

Common mistakes

1. Using a single blanket changeover duration instead of per-pair directional times. A single cleaning window applied to all transitions on a machine ignores the asymmetric contamination risk between SC and WG versus WG to SC, or between SC and EC versus EC to SC. The schedule will over- or under-estimate changeover time depending on the actual job sequence. Fix: Enter each from-to pair with its duration separately on the machine detail page — four entries per shared machine (two directions for each of the two shared class combinations) where both classes visit that stage.

2. Creating one product class for all liquid formulations. SC and EC share the same unit (litres) and both route through liquid filling, but their upstream paths are completely different — SC needs pre-mix and wet milling; EC skips both and enters at compounding. A single class forces every liquid product through the same routing, either creating phantom milling operations for EC or omitting them for SC. Fix: Create separate classes for SC and EC, each with its own routing, even though they converge at liquid filling.

3. Modelling Granulation and Drying as two separate stages. The dataset has two machines — an extruder granulator and a fluid-bed dryer — but they run sequentially per batch, not as independent parallel stages. Splitting them into two stages would allow the algorithm to interleave WG batches in ways that do not reflect the physical process (e.g., starting a new batch on the granulator while the previous batch is still in the dryer, when in reality the same physical batch occupies both machines serially). Fix: Model both machines as part of a single batch stage (Granulation & Drying). The stage duration for one full batch is the granulator's 360-minute cycle followed by the dryer's 240-minute cycle — the longer of the two per-batch times (here both are needed sequentially, so effective time is 600 minutes per batch).

4. Adding every SKU rather than one representative per class. A full product catalogue can run to dozens or hundreds of SKUs across three classes. Loading every variant multiplies setup effort without improving scheduling quality — products in the same class share routing and machine parameters, so additional SKUs add no new constraint information for the algorithm. Fix: Add one representative product per class (a typical high-volume SKU). Additional products within the same class can be added later without changing routing or machine configuration.

5. Forgetting to revert the seasonal calendar after the peak window ends. If the peak-season calendar is set permanently instead of on a date range, the schedule continues to use double-shift hours year-round, overstating available capacity during off-peak months and making the Gantt show unrealistic completion times. Fix: Assign the Peak Season calendar to the mid-January to mid-April date range via a schedule calendar period. The Standard default calendar covers all dates outside that range automatically.

What a good schedule looks like

With the configuration described in this guide, the scheduling algorithm produces a plan that coordinates all three product classes through their divergent routes across the shared machines. Before running the schedule against Schantt's optimisation, a typical manually built plan in a spreadsheet or whiteboard session shows these symptoms:

Before (manual or spreadsheet planning):
- Changeover sequences are ordered by planner intuition, not minimised — a batch sequence of SC, WG, SC on a disperser incurs two long cross-class cleanings (SC to WG at 135 minutes, then WG to SC at 90 minutes) when a grouped sequence (SC, SC, WG) would incur only one
- The QC hold for a batch finishing late on Friday is manually padded to account for the weekend, a step that is easy to forget or mis-estimate, leaving the filling line idle Monday morning
- The wet mill queue alternates between SC and WG jobs arbitrarily, causing unnecessary direction changes and their associated cleaning windows
- Machines on the Standard calendar are shown as available during off-peak hours in the peak season, or vice versa, because the calendar switch is applied inconsistently
- The planning team spends several hours each week re-sequencing jobs, adjusting for missed transitions, and reconciling downstream start times against the actual calendar

After (Schantt Auto mode):
- The algorithm sequences jobs to favour grouped runs within the same class, reducing the number of cross-class changeovers and cutting the total time spent on cleaning transitions
- The QC hold transfer time extends automatically across non-working days — a Friday batch shows as available for filling after the weekend has elapsed, without manual padding
- Wet mill jobs are balanced across both bead mills, with partial transfers allowing compounding or granulation to begin once the first 1,000 kilograms clears the mill, compressing the overall timeline
- The Peak Season calendar applies automatically from mid-January to mid-April; the Standard calendar takes over outside that range, reflecting actual available capacity without planner intervention
- The planning team reviews and adjusts the Gantt — typically in a fraction of the time previously spent on manual sequencing — and can experiment with what-if scenarios by adding or reordering jobs