Examples of use cases that custom simulations can solve
Custom simulations can solve many network optimisation use cases. The platform leverages multiple optimisation models under the hood, the data gathered from the audit product as well as additional data pipelines.
Milk-round optimisation
If demand is fulfilled to multiple destinations from stocking or supply points it is often optimal to drop goods off in a milk-round i.e multiple destination locations for one shipment. The objective of the optimisation is to find the optimal routing and order of drop off to maximise some outcome.
Network
60 destination locations
4 logistics providers
1000 SKUs
10 rate cards
10 manufacturing locations
Optimisation Target
Minimising cost of fulfilment
Model & Constraints
At a high-level this model optimises over the following dimensions:
- When to collect and deliver each purchase order, given an allowable “flex” in each date, and full route time estimates with site opening times
- Consolidation of orders, quantified as number of pallets, into vehicles
- Exact loading configuration i.e. which consignment is transported by which vehicle, given certain constraints
- Vehicle route taken; both direct routing and “milk-run” routing are considered with up to 5 sites visited including up to 2 delivery sites
- Primary/secondary favoured LSPs
Within each Experiment, we run several different simulations with different constraints to understand the impact on cost of varying factors such as the flex in dates, loading configuration rules, whether to use milk-run routing, etc
Outputs
A set of shipment recommendations where the shipment includes:
- when goods should be picked up
- from what location
- with provider, contract and service/vehicle type
- what destinations should be shipped to
- the order goods should be dropped off
Emissions optimisation
Customers try to reduce the CO2 emissions impact of their logistics operations. We have models that produce recommendations on what shipments and how to send them to minimise their carbon footprint.
Network
3 supplier locations
2 destination warehouses
1 logistics provider
1 rate card
Optimisation Target
Minimise shipment CO2 emissions
Model & Constraints
At a high-level this model optimises over the following dimensions:
- Mode
- Routing
- Vehicle type
- Fuel type
Each experiment runs multiple scenarios based on different levels of constraints on the above variables as well as constraints on due dates of each of the demands. The different simulations represent the possible emissions reductions depending on how much flexibility the carrier has to change the dimensions mentioned above.
Outputs
A set of shipment recommendations that include:
- Carrier
- Ship date
- Origin port
- Destination port
- Pre leg/main leg/post leg
- mode
- vehicle type
- fuel type
- Estimated arrival time
Supply optimisation
Logistics optimisation is often combined with supply optimisation in order to get recommendations on how much and what stock to purchase as well as how to ship it.
Network
5 suppliers
10 logistics providers
20 rate cards
10 stocking locations
Optimisation Target
Cost reduction where cost is a function of: transport cost, loss of sale, cost of storage, cost of write-offs
Model & Constraints
The model optimises what SKUs to order and how to order them as well as what logistics provider and service/mode to use to fulfil this order. At a high-level the following constraints are considered:
- Supplier information - location, inventory limits
- SKU ordering limitations - MOQs, order multiples, price, cost
- Provider rate cards - lanes, services, modes, load types
- Order target - due dates, goods in limits
- Routing constraints
Outputs
A set of recommended purchase orders and shipments or changes to existing purchase orders shipments if in place.
Stock balancing
This model considers no new supply into the network but figures out how to balance existing stock across locations in an optimal way
Network
25 locations
10000 SKUs
4 logistics providers
20 rate cards
Optimisation Target
Transportation, inventory cost reduction, stock out reduction
Model & Constraints
The transfer model balances cost of transportation and storage with stock outs. It takes the following constraints and parameters into account:
- Demand forecast - SKU, time, location demand forecasts
- Supplier information - location, inventory, inventory limits
- Provider rate cards - lanes, services, modes, load types
- Routing constraints
Outputs
A set of recommended transfer orders and shipments to move stock between a network
Advanced procurement simulation
Procurement analyses can be made more complex by the inclusion of volume based rebates and the desire to have primary and secondary carriers on each lane for redundancy purposes. This creates a more complex optimisation problem than the standard procurement simulation template.
Network
2 locations
4 providers
4 rate cards
500 lanes
Optimisation Target
Model & Constraints
This model optimises over:
- Where to fulfil each shipment (order) from
- In this case we are choosing from Distribution Centres in the UK or Ireland and fulfilling sales orders internationally
- Which provider to use for each shipment
- Minimising total cost, as a combination of immediate shipment cost, and provider rebates
- Rebates calculated over all global shipping spend, using qualifying spend to identify shipment tier, with percentage applied to rebatable spend
- Ensuring operational resilience: Each destination region is required to have 2 providers
- A primary with at least 60% and a secondary with at least 20%
- Ensuring procurement efficiency and operational efficiency:
- Only 2 providers per region
- Only 1 provider per city
Outputs
A set of shipments and the primary and secondary logistics provider awarded with the volume on that lane.
Other
The above are just examples of how we can apply our simulation technology to solve logistics and supply chain optimisation problems.