Case studies for the curious…

Courier - A need for Speed

Planning in the courier space is notoriously difficult.

For one, the sheer number of jobs and hence locations to route to are significantly higher than usual. While the Icepack scheduler is able to perform rapid turnarounds, once order sizes reach multiples of 10,000’s and fleet sizes are greater than 100 it makes sense to try and decompose the problem prior to solving.

Chalcid found significant improvements over current operations by using a combination of the natural decomposition of orders by region, automatic order aggregation, and searching for optimal service areas given depot locations.

Savings of 10% with respect to distance travelled and 25% reduction in overall cost where discovered. It should be noted that the study shows how savings can be made as planners move from master routes to dynamic scheduling. In practice dynamic scheduling at this scale is difficult to to achieve due to complexities that arise between vehicle loading and intra-warehouse constraints.

Waste Collection - Keep it clean

Wast collection planning is significantly different to normal delivery and pick-up operations.

An important aspect of the model is understanding how to translate the SLA requirements, consolidate successive jobs along road segments and correctly identify the required underlying network locations to pass to Icepack’s Street Router solver which in turn caters for contiguous jobs along a road segment, vehicle orientation and side of road constraints.

Chalcid has developed a number of techniques which have aided the process of building sufficiently accurate models using available location and road-network data to provide insight into the work loads and schedules that should be performed across different planning horizons.

Dairy - Not just milk

Large scale network analysis using Icepack’s Network Source Model to determine optimal primary movements of SKU’s between primary and secondary depots. In addition to this the output of the Network sourcing was fed into Icepacks Vehicle Routing solver which allowed us to asses the impact on the secondary distribution footprint.

This allows us to consider the impact of secondary distribution costs while solving line-haul and primary stock movements between warehouses, depots and satellite depots. Optimising 50+ product types across 10 producers, 20 primary depots and 3000 customers.

Our ability to rerun multiple scenarios as and when needed, incorporating updated information, and what-if conditions, allows primary and secondary strategies to be analysed and performed as required and get results to decision makers quickly.

Order Consolidation - Hold it together

Consolidation study of 4 large distribution footprints comprising of a total of 30 000 orders was achieved using Icepack’s IVR solver in conjunction with order analysis by Chaclid.

The orders across the footprints where consolidated down to 8 000 orders which represented overlap between the footprints. Next, automatic clustering of orders that should have been group together which have common delivery points were performed reducing the final orders set to around 4 000.

Considerably savings can be found when consolidation is possible and warehouse picking systems can be aligned with new regional delivery patterns, either as master routes or dynamic dispatch.

Fleet Configuration - Go big or Go small

What impact will changes to a fleet configuration have on distribution costs, utilisation, and delivery times?

How will future demand affect fleet requirements?

What is the impact of adding new vehicles, changing the mix of vehicles, or reducing the fleet either in practice or to assess the excess capacity available in case of unforeseen events.

In one study we reviewed the fleet configuration for a problem with 7 000 locations, 22 000 orders, 3 depots and 4 vehicle types. The problem also had a variety of time-window and specific delivery cluster requirements for which efficient routing and vehicle mixes had to be found.

Street Sweeper Routing - Plan a clean getaway

Mechanical sweeping is a vital part of keeping city centers looking clean, removing harmful materials, and reducing storm water pollution. Chalcid performed data and network analysis by segmenting approximately 2 500 roads into 10 000 segments that could then be modelled using Icepacks ISR solver.

The area, covered by 4 mechanical street sweeping vehicles of 2 distinct types, required adherence to an SLA which ranged from daily, quarterly and year cleaning activities across the road-network. 

Initial bench marking found up to 20% savings in total distance travelled when keeping all tasks to be performed on a day the same, and almost 50% saving if ad-hoc tasks are planned. While this represents a hypothetical best-case scenario, where there are no unexpected jobs to cater for, it still shows the impact ad-hoc work has compared to when it is planned.