Scenarios and Pipelines

Route: /pipelines

The Pipelines screen is the command centre for managing forecast runs, inventory optimisation scenarios, supply planning runs, and workflow execution. It is organised into five tabs:

Tab	Icon	Purpose
Forecast	🔀	Manage forecast scenarios and run the forecasting pipeline
EOQ	📐	Configure and view Economic Order Quantity curves
MEIO	📦	Multi-Echelon Inventory Optimisation scenarios and results
Supply Planning	🚛	Supply projection scenarios
Pipeline	▶️	Workflow orchestration — run multiple steps in sequence

---

Understanding Pipelines and Scenarios

What is a Scenario?

A scenario is a named set of parameters and outputs for one planning run. Every pipeline step (forecasting, MEIO, supply planning) stores its outputs under a scenario ID. This allows you to keep multiple versions of the forecast or inventory plan simultaneously:

Figure: the families of scenario types managed across the Pipelines tabs.

mindmap
  root["Scenario types"]
    Forecast
      Base scenario
      High demand
      Long lead time
    MEIO
      Base scenario (is_base = true)
      Per-segment overrides
    Supply
      Supply scenario
        own supply_scenario_id
    Causal
      Fleet deployment
      Failure rates
    Maintenance
      "Planned maintenance"

• Base scenario — the current agreed plan
• High demand scenario — what if demand grows 20%?
• Long lead time scenario — what if supplier lead times increase by 4 weeks?

Each scenario has its own set of results in the database. Switching between scenarios in the left sidebar pipeline selector instantly changes the data shown across all screens.

What is a Pipeline?

A pipeline is a named execution context that links together scenarios from different steps. A pipeline has:

Figure: how a pipeline links a base scenario of each type into a single execution context.

erDiagram
  PIPELINE ||--o| FORECAST_SCENARIO : "links"
  PIPELINE ||--o| MEIO_SCENARIO : "links"
  PIPELINE ||--o| SUPPLY_SCENARIO : "links"
  PIPELINE ||--o| CAUSAL_SCENARIO : "links"
  PIPELINE ||--o| MAINTENANCE_SCENARIO : "links"
  FORECAST_SCENARIO ||--o| BASE_FORECAST : "has base (is_base=true)"
  MEIO_SCENARIO ||--o| BASE_MEIO : "has base (is_base=true)"
  PIPELINE {
    string name
    int pipeline_id
  }
  FORECAST_SCENARIO {
    string name
    bool is_base
  }
  MEIO_SCENARIO {
    string name
    bool is_base
  }
  SUPPLY_SCENARIO {
    string name
    int supply_scenario_id
  }

• A name (e.g. "Production Run Q2 2026")
• A linked forecast scenario
• A linked supply scenario (with its own supply_scenario_id)
• Optionally a linked MEIO scenario and causal scenario

When you select a pipeline from the sidebar dropdown, all screens use the scenario IDs associated with that pipeline to retrieve results.

The Pipelines screen lets you manage forecast, MEIO and supply scenarios.

The pipeline is the unit of a planning cycle

Each weekly or monthly planning cycle typically corresponds to one pipeline run. You run all the steps (ETL → Outlier → Characterisation → Forecasting → MEIO → Supply), and the outputs are stored under the pipeline's scenario IDs. Planners then review and approve the results.

Immutable vs Pipeline-Scoped Data

Data	Pipeline-scoped?
Items (item master)	No — shared across all pipelines
Sites	No — shared
Demand actuals	No — shared (historical data is not re-written by running a pipeline)
Firm orders	No — shared
Forecast results	Yes — stored per scenario
MEIO results	Yes — stored per scenario
Supply projections	Yes — stored per scenario
Exception records	Yes — stored per pipeline
Forecast overrides	Yes — stored per pipeline
IO overrides	Yes — stored per pipeline/scenario

This design means you can run alternative scenarios without touching historical data or master data.

---

Forecast Tab — Forecast Scenarios

The Forecast tab (ScenarioPlanner) shows all forecast scenarios.

Creating a Forecast Scenario

1. Click + New Scenario.
2. Enter a name and optional description.
3. Optionally mark it as the Base scenario.
4. Save.

The new scenario appears in the list. It has no results yet — you need to run the forecasting step to populate it.

Running Forecasting

Select one or more scenarios using the checkboxes, then click Run Forecasting. The system submits a background job. Job status updates automatically (Running → Completed or Error).

After completion, navigate to the Dashboard and select this scenario's pipeline from the sidebar dropdown to see the results.

Scenario Comparison

The Forecast tab allows side-by-side metric comparison across scenarios. Select multiple scenarios and the comparison chart shows accuracy metrics (MASE, MAE, Bias) per scenario, aggregated across all series.

---

EOQ Tab — Economic Order Quantity

The EOQ tab (KCurve) shows the Economic Order Quantity calculation for all items.

EOQ is the order quantity that minimises the total cost of ordering and holding inventory:

EOQ = √(2 × Demand × Order Cost / Holding Cost per unit)

The K-curve view plots the total cost as a function of order quantity, showing the cost-minimising EOQ.

Running EOQ

EOQ requires the following inputs per item:

• Annual demand (from forecast or historical data)
• Order cost (configured in Settings or item master)
• Holding cost per unit (from Settings)

Click Run EOQ to compute EOQ for all items in the active scenario. Results are stored per scenario.

Using EOQ

After running, EOQ values appear in the Exceptions IO Cell for items that have an IO exception. Planners can override the computed EOQ directly from the exceptions screen.

Order-frequency constraints in EOQ

When an order_frequency parameter is configured (see Parameters), each item's EOQ can be snapped to the nearest allowed replenishment frequency. The K-Curve workbench has an "Apply order-frequency constraints" toggle to enable constrained mode. When active:

• The exchange curve shows an orange stepped overlay representing the constrained frontier.
• Feasible k alternatives appear as clickable buttons, letting you jump between operating points.
• Item rows with frequency constraints expand to show all allowed frequency options (Q, orders/year, total cost).

In the per-item EOQ detail view, a new "Order Frequency Options" panel lists each allowed frequency with its cost, and orange markers on the cost curve highlight where each allowed quantity sits.

---

MEIO Tab — Multi-Echelon Inventory Optimisation

The MEIO tab (MeioScenarios) manages inventory optimisation scenarios. MEIO (Multi-Echelon Inventory Optimisation) computes the optimal safety stock level for each item-site that maximises service level subject to budget or other constraints.

See the Supply Planning guide for how MEIO results feed into the supply projections.

MEIO Scenarios

A MEIO scenario contains:

• Global parameters — apply to all items unless overridden
• Per-segment parameters — override global parameters for specific segments

The parameter table shows one row per segment (plus a Global row). Columns are parameters; cells are inline-editable.

MEIO Parameters

Parameter	Label	Description
fill_rate_target	FR Target	Target line-fill-rate for this group (0–1). Example: 0.95 = 95%
max_budget	Max Budget	Maximum total inventory investment (blank = unlimited)
sku_min_fill_rate	SKU Min FR	Minimum allowed fill rate per individual SKU
sku_max_fill_rate	SKU Max FR	Maximum allowed fill rate per individual SKU
demand_multiplier	Demand ×	Multiplier applied to all demand rates (e.g. 1.2 = assume 20% more demand)
lead_time_multiplier	LT ×	Multiplier applied to replenishment lead times
sku_min_sl_qty	Min SS Qty	Minimum safety stock quantity in units
sku_max_sl_qty	Max SS Qty	Maximum safety stock quantity in units (blank = no cap)
sku_min_sl_slices	Min Cover	Minimum months of demand cover
sku_max_sl_slices	Max Cover	Maximum months of demand cover (blank = no cap)
holding_cost_pct	Holding %	Annual holding cost as % of inventory value
use_existing_inventory	Use Inv.	If on, credits on-hand stock against the recommended safety stock
return_rate	Return %	Fraction of demand returned (overrides RETURN route value)
repair_yield	Repair %	Fraction of returned units successfully repaired
line_fill_rate	Line Fill Rate	(Global only) Use line-fill-rate mode vs unit-fill-rate
consider_eoq	Consider EOQ	(Global only) Use computed EOQ as the order quantity

Editing MEIO Parameters

Click any cell in the parameter table to edit it inline. The cell switches to a number input. Press Enter to commit, Tab to commit and move to the next cell, or click away to commit. Changes take effect on the next MEIO run.

Values shown in muted italic are inherited defaults — the global value is being used because no per-segment override is set. Click an inherited value to set an explicit override for that segment.

All number values follow the Number of Decimals setting from Settings.

Running MEIO

After configuring parameters, click Run MEIO for the selected scenario. MEIO is computationally intensive and runs as a background job. For large datasets (hundreds of SKUs), expect several minutes.

Results appear in the results table: for each item-site, the recommended safety stock quantity, the expected fill rate, and the implied service level.

---

Supply Planning Tab

The Supply Planning tab (SupplyScenarios) manages supply projection scenarios. See Supply Planning for full details.

---

Pipeline Tab — Workflow Orchestration

The Pipeline tab lets you run a complete sequence of pipeline steps in one click, and review the run history.

Creating a Workflow Pipeline

A pipeline workflow links together:

1. Forecast scenario to use
2. Which steps to run (ETL, Outlier Detection, Characterisation, Forecasting, Evaluation, MEIO, Supply, etc.)
3. Which items or segments to process

Click + New Pipeline to create one, then configure the steps.

Running a Pipeline

Workflow execution status and run history are shown in the Pipeline tab.

Figure: lifecycle of a pipeline workflow run as shown in the Pipeline tab.

stateDiagram-v2
  [*] --> Created: new pipeline configured
  Created --> Queued: planner clicks Run
  Queued --> Running: steps execute in order
  Running --> Completed: all steps succeed
  Running --> Failed: a step errors
  Failed --> Created: fix & re-run
  Completed --> [*]: results stored under scenario IDs
  Completed --> Reviewed: planners review results
  Reviewed --> Created: next planning cycle

Select a pipeline and click Run. The steps execute in the configured order. Progress is streamed live to the screen: each step shows its start time, end time, duration, and any log messages.

When run as part of a pipeline workflow, each step uses the pipeline's own scenario IDs — not the scenario from the sidebar dropdown. This is important for automation: workflows are self-contained and do not depend on what a planner has selected in the sidebar.

Pipeline History

The bottom of the Pipeline tab shows a log of recent pipeline runs: when they ran, which user triggered them, how long they took, and the final status (completed / failed).

---

The Sidebar Pipeline Selector

The pipeline selector dropdown at the top of the left sidebar sets the active pipeline for all interactive screens. When you change the selected pipeline:

• The Dashboard refreshes to show that pipeline's forecast results
• The Exceptions screen reloads with that pipeline's exceptions
• The Time Series Viewer loads that pipeline's forecasts and supply projections
• The MEIO and Supply tabs show that pipeline's results

The pipeline selector is for interactive review. When steps run as part of a workflow, they use the workflow's pipeline — not the sidebar selection.

---

Multi-Tenancy

Each tenant (organisation) has their own isolated database with their own pipelines and scenarios. If your user account has access to multiple tenants (common for system administrators managing several clients), you can switch between them using the account switcher at the bottom of the sidebar. Switching accounts logs you into a different tenant database entirely.