Production Scheduling for Fruit & Vegetable Canning

If you plan production for a fruit and vegetable canning facility, you juggle overlapping harvest campaigns, pH-regime changeovers between product classes, and a retort bottleneck that can make or break a season. This guide shows how to model your canning line in Schantt, configure it stage by stage, and produce optimised schedules that respect your process constraints without manual spreadsheet wrangling.

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

Industry context

Fruit and vegetable canning begins with raw produce arriving in seasonal surges: peaches ripen in mid-summer, sweet corn and tomatoes follow in late summer, and each crop must be processed within a narrow quality window before it spoils. A mid-market canning facility receives, washes, grades, blanches, fills and seams, retorts, cools, and labels multiple product classes across a seven-stage production line. Each stage presents its own constraint — the retort stage, typically four rotary batch vessels, is the bottleneck operating at 85–95% utilisation during peak pack. The filling stage must switch between product classes with either a quick rinse (same pH regime) or a full clean-in-place cycle (cross-regime), and the blancher upstream must supply material fast enough to keep the filler running.

Changeovers between low-acid and high-acid products require thorough cleaning that consumes 45–60 minutes per switch, while same-regime switches need shorter rinses. These changeover durations compound when the planner has to sequence multiple campaigns across the same machines and cannot afford to waste a single shift of available capacity.

Facilities in this segment process around 9,000 tonnes of finished product annually, with raw throughput reaching 18–24 tonnes per eight-hour shift at the seasonal peak. The plant runs three overlapping campaigns — peach, corn, and tomato — compressed into roughly 16 peak weeks from July through October.

Valley Harvest Canning Co. runs ~120 people (200 at peak) at a 4,500 m² facility, making 3 product classes across 7 production stages, scheduled by a 3-person planning team.

Process overview

flowchart LR
    S1["Receiving & Washing<br/>Flow"]
    S2["Grading & Sorting<br/>Flow"]
    S3["Blanching<br/>Flow"]
    S4["Filling & Seaming<br/>Flow"]
    S5["Retorting<br/>Batch"]
    S6["Can Cooling<br/>Flow"]
    S7["Labelling & Packaging<br/>Flow"]
    S1 --> S2
    S2 --> S3
    S3 --> S4
    S4 --> S5
    S5 --> S6
    S6 --> S7

Seven production stages move raw produce from receiving through washing, grading, blanching, filling and seaming, retorting, cooling, and finally labelling and packaging.

Product-class routing note: Corn Products skip Grading & Sorting. The kernel corn route bypasses the size-grading belt, passing directly from Receiving & Washing to Blanching via a flume bypass. Peach and Tomato Products traverse the full stage sequence.

Scheduling challenges and how Schantt handles them

The schedule is driven by harvest campaigns — the available tonnage of each crop and the date window before quality degrades. If your plant runs on customer-order pull or a mix of harvest and contract-pack, the same modelling applies: you set the scheduling horizon and let the system optimise against it. Schantt always schedules forward from a start date and minimises the total production time needed to complete the work. For canning, the practical horizon is the full peak season — typically 16 weeks — so the system can span the overlapping campaigns in a single run.

Schantt offers two optimised scheduling modes. Auto mode explores both the job sequence and machine assignments to find the arrangement that minimises total production time. Semi-Auto mode lets the planner lock the job sequence and lets the system optimise machine assignments around it — useful when you need to enforce a strategic order (process the highest-value crop first).

What Schantt handles well

• Sequential multi-stage production with stage skipping — Schantt models the full canning line as an ordered stage sequence; product classes that skip intermediate steps use per-class routing with bridging transfer times so handoff delays are preserved.

• Multiple parallel machines per stage — Retorts run as parallel batch machines; fillers and blanchers run as parallel flow machines. The scheduling algorithm explores machine assignments across each stage to find the combination that minimises total production time.

• Mixed batch-and-flow processing in one route — Canning mixes flow stages (washing, filling, labelling) with a batch stage (retorting). Schantt handles both in a single route: flow stages use throughput rate, batch stages use cycle time and batch size. The simulation walks each job through its mixed-type stages.

• Sequence-dependent changeovers — Changeovers in canning vary by product-class pair: a rinse between same-regime runs takes minutes, a full CIP between different vegetables takes much longer. Schantt models these as directional per-machine entries; the optimiser naturally clusters similar products to reduce total changeover time.

• Simulation-based material supply tracking — When a downstream stage outruns its upstream supply, Schantt inserts wait-material pauses on the starved stage. The planner sees these as explanatory gaps on the Gantt.

• Calendar-aware scheduling with seasonal transitions — Off-season single shift, shoulder double shift, and peak triple shift — each with different working hours and days — are modelled as named calendars. Schedule calendar periods let the planner assign each date range its calendar so the schedule respects the right shift pattern automatically.

How Schantt handles each challenge

1. Retort bottleneck and pH-regime changeovers.

• The four retorts run near capacity during peak pack — a rinse between same-regime runs takes 35 minutes, but a full clean-in-place between different pH regimes takes 45 minutes. Every cross-regime swap costs three quarters of an hour of lost throughput on the bottleneck stage. With manual scheduling, a planner may optimise one retort's sequence while another sits idle waiting for a batch.
• Schantt models each retort as a batch machine with per-class cycle time and batch size — 22 minutes at ~1,100 kg for peach, 38 minutes at ~1,000 kg for corn, 30 minutes at ~1,200 kg for tomato. Each product class has a regulatory thermal process (time and temperature validated by a process authority), and Schantt stages retort cycle times to match these validated durations. The scheduler sequences jobs across all four retorts simultaneously, respects the directional changeover durations the planner enters, and naturally groups same-regime runs to minimise cross-regime swaps.

2. Filler changeover sequencing across product classes.

• The consumer filler runs at 250 cans per minute and the institutional line at 80 cans per minute. Switching between corn and a fruit product requires a 55–60 minute clean-in-place; switching within the same pH regime (peach to tomato) takes 40–45 minutes. Without deliberate sequencing, these changeover windows fragment the day, especially when the planner must switch between campaigns multiple times a week.
• Schantt models changeovers as directional per-machine entries — the planner enters the actual duration for every from-product-class to to-product-class pair on each filler. The scheduling algorithm sequences jobs to minimise total changeover time consumed on each filler, clustering similar products naturally.

3. Overlapping harvest campaigns compressed into 16 peak weeks.

• Three campaigns — peach (~2,100 t), corn (~2,500 t), and tomato (~3,500 t) — overlap in August and September. The planner must judge whether the full tonnage fits before each crop's quality window closes, relying on rough spreadsheet estimates and experience. A scheduling guess that underestimates changeover durations by even five minutes per switch can cascade into lost throughput across the season.
• The planner sets the schedule start and end dates to bound each campaign. Schantt schedules forward from the start date, and the Gantt shows whether forecast tonnage completes before the end date. The planner gives the scheduler a realistic horizon — the full 16 peak weeks — and verifies fit visually. If the Gantt shows a campaign tail extending past the window, the planner adjusts priorities or shifts machine assignments and re-runs.

4. Material starvation between upstream supply and downstream throughput.

• The blancher feeds the filler at up to 6 t/hr, but the filler alone can draw 6.4 t/hr for consumer cans or the equivalent of 14 t/hr for institutional cans. During peak season this mismatch causes 2–4 hours of unplanned line idle per week when the filler runs ahead of supply, and the effect cascades when retort crews wait for filled cans to arrive.
• Schantt simulates each job through its stages sequentially. When a downstream stage draws material faster than its upstream supply, the scheduler inserts wait-material pauses on the starved stage — these appear as explanatory gaps on the Gantt. The planner sees exactly where the line starves and can adjust job sequences or shift the production order to keep the blancher continuously feeding the filler.

What to model in Schantt

To represent a fruit and vegetable canning facility in Schantt, you configure a set of first-class entities that mirror the physical plant and its product range.

Step-by-step setup

1. Create the seven stages in order. Open the Stages page and add each production stage at its correct position: Receiving & Washing (position 1), Grading & Sorting (2), Blanching (3), Filling & Seaming (4), Retorting (5), Can Cooling (6), Labelling & Packaging (7). Set each stage's production type — flow for all except Retorting, which is batch. On each Stage detail page, configure the transfer times between consecutive stages. Key values:
- Receiving & Washing → Grading & Sorting: 10 minutes
- Grading & Sorting → Blanching: 8 minutes
- Blanching → Filling & Seaming: 12 minutes
- Filling & Seaming → Retorting: 5 minutes
- Retorting → Can Cooling: 15 minutes
- Can Cooling → Labelling & Packaging: 10 minutes

Add the bridge transfer for the Corn skip route: Receiving & Washing → Blanching at 6 minutes.

2. Add the sixteen machines to their stages. On the Machine page, create each machine and assign it to the correct stage — two receiving washers (RW-01, RW-02), two grading sorters (GS-01, GS-02), two blanchers (BL-01, BL-02), two filler-seamer lines (FS-01, FS-02), four retorts (RT-01 through RT-04), one cooling canal (CC-01), and three labelling-and-packaging machines (LP-01, LP-02, LP-03). The machine count per stage must match the physical floor — over- or under-counting changes the bottleneck behaviour.

3. Create the product classes and configure routing. Define three product classes: Peach Products (high-acid, full routing through all seven stages), Corn Products (low-acid, skips Grading & Sorting), and Tomato Products (acid, full routing). On each Product Class detail page, set the per-class routing:
- Peach routing: Receiving & Washing → Grading & Sorting → Blanching → Filling & Seaming → Retorting → Can Cooling → Labelling & Packaging
- Corn routing: Receiving & Washing → (skip Grading & Sorting) → Blanching → Filling & Seaming → Retorting → Can Cooling → Labelling & Packaging
- Tomato routing: same full sequence as Peach

Enable the partial-transfer toggle on Filling & Seaming for all three classes and set the transfer quantity to 130 kg — this lets a filler basket start moving to retort before the full batch is complete, mimicking real line overlap.

4. Add the three products. Create one product per class: Peach halves in heavy syrup (Peach Products), Whole kernel corn in brine (Corn Products), Diced tomatoes in juice (Tomato Products). Assign each to its class and give it a display colour for the Gantt.

5. Set machine capacity parameters and changeovers. For each machine on the Machine detail page, configure the relevant capacity parameters. Flow stages take throughput values in units per hour — these represent the machine's run rate for each product class that reaches it. The Retorting stage, being batch, takes cycle duration and batch size per basket load.

Key capacity values:
- Receiving & Washing throughputs: RW-01 at 8 t/hr (peach, tomato) or 7.5 t/hr (corn); RW-02 at 6 t/hr (peach, tomato) or 5.5 t/hr (corn)
- Blanching throughputs: BL-01 at 5.5 t/hr (peach), 6 t/hr (corn), 5.8 t/hr (tomato); BL-02 at 3 t/hr (peach), 2.8 t/hr (corn, tomato)
- Retort batch parameters: Peach cycle 22 min at ~1,100 kg per load; Corn cycle 38 min at ~1,000 kg; Tomato cycle 30 min at ~1,200 kg — values are identical across all four retorts
- Filler-seamer throughputs: FS-01 at 6.4 t/hr (peach, tomato) or 6.2 t/hr (corn); FS-02 at 14 t/hr (peach, tomato) or 13.5 t/hr (corn)
- Cooling canal throughput: CC-01 at 7.2 t/hr (peach, tomato) or 7 t/hr (corn)

Changeovers are configured per machine, per direction — the actual duration for every from-product-class to to-product-class pair on each machine that sees more than one class. Enter these directional entries:
- Filler changeovers (FS-01, FS-02): cross-pH-regime transitions (corn to or from peach or tomato) at 55–60 minutes; same-regime transitions (peach to or from tomato) at 40–45 minutes
- Retort changeovers (RT-01 through RT-04): cross-regime 45 minutes, same-regime 35 minutes, for every pair
- Blancher, receiving, grading, cooling, and labelling changeovers: shorter durations of 5–20 minutes depending on the machine and the product-class pair

6. Configure calendars, exceptions, and downtimes. Create three calendars — Standard (off-season, Monday to Friday 08:00–16:30), Shoulder (Monday to Saturday 06:00–22:00), and Peak-Harvest (Monday to Saturday 06:00–24:00 plus Sunday 06:00–14:00). Set Peak-Harvest as the default. Add calendar exceptions for New Year's Day, International Workers' Day, and Christmas Day. Add machine downtimes: RT-01 annual inspection in February, FS-01 pre-season filler overhaul in late May, and the year-end facility shutdown from 24 to 31 December.

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

Common mistakes

1. Using one blanket changeover time instead of per-pair durations. Entering a single changeover value for all transitions on a filler or retort ignores the difference between a same-regime rinse and a full pH-regime clean-in-place. The scheduler cannot distinguish fast transitions from slow ones, so it misses opportunities to cluster similar products. Fix: Enter the actual directional duration for every product-class pair on each machine — 45 minutes for a cross-regime retort changeover versus 35 minutes for same-regime.

2. Creating one product class covering all three pH regimes. A single class that bundles peaches, corn, and tomatoes loses the divergent routing for Corn (skip Grading & Sorting) and the different retort cycle times. The scheduler treats all three as identical, producing an infeasible schedule. Fix: Create separate product classes for Peach Products, Corn Products, and Tomato Products — each with its own routing and per-class machine parameters.

3. Modeling retorts as flow machines instead of batch machines. Setting retort throughput in tonnes per hour instead of configuring cycle time and batch size loses the discrete batch physics — retorts process one basket load at a fixed duration, not a continuous stream. The scheduler may over- or under-estimate retort utilisation. Fix: Set the Retorting stage production type to batch and enter per-class cycle duration and batch size on each retort machine.

4. Ignoring the stage-skip routing for Corn Products. Defining a full seven-stage routing for Corn Products sends kernel corn through Grading & Sorting, which in reality it bypasses via a flume. The schedule includes unnecessary processing time and potentially blocks the graders for products that do need them. Fix: On the Corn Product Class routing, omit Grading & Sorting and add a bridge transfer time from Receiving & Washing directly to Blanching (6 minutes).

5. Setting all stages on one calendar without seasonal transitions. Using the same shift pattern for the full year misses the plant's fundamental operational rhythm — single shift off-season, double shift in shoulder, triple shift plus Sunday during peak pack. The schedule shows the wrong available hours for most of the year. Fix: Create three named calendars with their shift patterns, then use schedule calendar periods to assign each date range to the correct calendar.

What a good schedule looks like

A well-configured canning schedule shows clean campaign blocks flowing through all seven stages, with retort utilisation spread evenly across the four vessels, changeover time visible as bounded gaps, and material supply holding continuous across the blancher-to-filler handoff.

Before (manual planning): Planners build the schedule in spreadsheets, sequencing campaigns by crop-arrival date alone. Symptoms include:

• Retort utilisation varying widely — two vessels idle while one is overloaded — because machine assignments are decided independently of the job sequence.
• Changeover windows double-counted or omitted when filler and retort timings drift out of sync.
• Unexplained line idle when the filler runs ahead of blancher supply, discovered only when the line stops.
• The schedule re-done from scratch each time a crop-arrival date shifts, consuming 2–3 hours per what-if question.

After (Schantt Auto mode): With the facility modelled as described, the scheduler produces a complete campaign-season plan in a single run — weeks of overlapping production compressed into a single optimised timeline. The planner loads the seasonal tonnage, sets the start and end dates, and the system explores the sequence and machine assignments across all three campaigns simultaneously. Gains include:

• The Gantt shows each campaign as a contiguous block across the seven stages, with retort jobs assigned to the four vessels in parallel — no single vessel overloaded while others wait. The planner can verify at a glance that all four retorts share the load.
• Filler changeovers cluster naturally: peach-to-tomato transitions (45 minutes) appear where the scheduler grouped same-regime runs, while corn-to-fruit transitions (55–60 minutes) are fewer and purposeful. The total changeover time consumed across the season is the lowest possible for the given campaign mix.
• Wait_material gaps appear only where blancher throughput cannot keep pace — visible and explainable, not a surprise stoppage. The planner adjusts timing or shifts order to minimise the impact.
• What-if responses are a matter of adjusting parameters — crop arrival date, tonnage, machine availability — and re-running the scheduler, not rebuilding the entire spreadsheet from scratch. A question that used to consume half the morning resolves in minutes.