The customer is a major bulk logistics rail operator serving the New South Wales and Victorian agricultural supply chain.
The customer needed a more efficient way to schedule their trains across its network, from loading trains at grain silos and transporting bulk amounts of grain over hundreds of kilometres of rail network to unloading at domestic mills and terminals for export.
This rail value chain is large, complex and involves many stakeholders. The customer’s prime constraint is developing a schedule modelling constraints on network access caused by high volumes of rail traffic on regional, coastal, and metro lines owned and operated by multiple entities.
In addition to the limited track availability, many load and unload sites only allow access to rail services within narrow time windows. Rail plans must align to schedules provided by domestic and export terminals and servicing load points during daylight hours for safety purposes. Accounting for the availability of all stakeholders within a value chain while also producing an optimal rail plan presents a challenging problem to solve.
Before engaging Polymathian, the customer’s spreadsheet-based planning system was time-consuming and inflexible. In a dynamic environment where network availability is regularly updated, re-planning often continues until the plan is released. Rudimentary validation methods hinder the ability of planners to reschedule or evaluate other options.
As an existing RACE user for other commodities, the customer was keen to see how the tool would impact their grain operations.
The first step was to model their complex rail network, producers, rolling stock, and associated constraints within RACE. By capturing this data in the tool, planners are no longer responsible for manipulating vast amounts of information. RACE stores this data, using it to produce optimised plans in a fraction of the time it took previously.
RACE introduces consistency and optimisation into the planning process. Whereas previously, the quality of a plan was subject to the planner’s experience and rules of thumb, without optimisation for time, cost, asset utilisation, and any number of business goals.
Replace subjectivity and human decision making inherent in manual processes with data-driven and objective mathematically optimised planning.
Latent capacity identificationOptimal plans maximise performance with minimum resources.
Ad hoc pathingPaths through the network can be found for new and above-contract customer requests.
Streamlined planning processPlans produced in minutes, rather than days, allows personnel to focus on more strategic opportunities.
Contract mix optimisationWithin the model, contracts are evaluated for yield and impact on asset utilisation.
The planning module provides mathematically optimal answers for questions such as: