Andreas T. Ernst

Industry & Research Experience

Andreas Ernst completed a PhD in network optimisation at the University of Western Australia in 1995. He has spent 20 years working at CSIRO on developing innovative operations research solutions for a range of operations research applications. He is continuing his research in many of these areas including logistics, rostering and network optimisation at Monash University.

Andreas’ research has resulted in over 60 papers that have been cited over 4000 times with an h-index of 24. For a complete list see https://scholar.google.com.au/citations?user=hSDlhWUAAAAJ&hl=en.

Darlrymple Bay Coal Terminal: Developed a range of simulation and optimisation models that addressed various aspects of decision making in the supply chain. This included simulation models to assist strategic decisions on capital investment and operating rules. These models covered the supply chain from the mine stockpiles, through railing to port stockpiles and shipping. Other models were developed to create long term (3 month) railing and shipping plans and to optimise the tactical (12 day) and operational (daily) schedules.

Recreational Vehicle Scheduling: The two largest campervan rental companies operating in Australia and NZ (THL Rentals and Apollo Motorhomes) use the DVASS scheduling software developed by Andreas and his colleagues on a daily basis. The software creates optimised schedules for 1000’s of vehicles moving across all of Australia and NZ over an 18 month time horizon. It trades of a range of soft and real costs relating to relocations, product substitutions (upgrades) and maintenance requirements. DVASS also provides a real-time rescheduling service to update the schedules as new bookings arrive or other changes need to be made due to unforseen events. Versions of this software have been in use for over a decade.

Hunter Valley Coal Chain: Working since 2006 with the Hunter Valley Coal Chain Coordinator, Andreas was involved in creating software to assist with the planning and operations of this supply chain. This included a capacity planning model that optimises the capacity expansion required to meet rising export throughput requirements. It produces an optimised portfolio of infrastructure expansion projects that ensure there are no bottle necks while minimising the overall cost. At the other end of the planning time scale, the Rail Scheduling Tool optimises the daily schedule of train trips [50]. It reduces the time taken to create rail schedules from a whole day of manual planning to just minutes. Andreas has been a partner investigator in two ARC grants lead by Natashia Boland at the University of Newcastle. These have focussed on issues around maintenance scheduling and stockyard management.

Rail Scheduling: In addition to the rail scheduling work for the Hunter Valley Coal Chain, Andreas Ernst has been involved in a number of other projects relating to rail scheduling. These include development of a rail crew rostering optimiser for National Rail.  Medium term rail planning software developed for RioTinto not only speeds up the planning time, but also produces solutions with increased throughput (0.5Mt worth approximately $100m in 2011). In another major project, track maintenance scheduling software has been developed that takes into account both the impact on track throughput and the constraints on maintenance crews. Simulation studies indicate a significant increase in throughput can be achieved with such smarter scheduling.

Rostering: Andreas Ernst was a key researcher involved in the development of an innovative toolkit of rostering engines for generating optimised demand responsive rosters. A novel feature of the approach to rostering developed at CSIRO was the use of a flexible rule engine that could be used to specify arbitrary hard rules, soft rules and employee preferences that should be considered in creating rosters. This feature, combined with the use of multiple underlying optimisation engines, allowed the rostering toolkit to be adapted to the wide range of rostering requirements encountered in businesses without needing to rewrite the core software. While initial attempts at commercialisation were unsuccessful, Andreas is currently negotiating a licensing arrangement for this technology.  Along the way the rostering tools have been used for a number of clients, particularly in strategic planning. Clients include the NSW Police, South Australian Ambulance Service, Australian Customs and others.

Hub Location: Andreas Ernst is an international expert in the modelling and solution of hub location problems. These network design problems arise in a variety of applications where flow between locations is routed via hubs or collection & distribution centres. Over a number of years Andreas has contributed to research into new algorithms for solving larger scale problems in this area and to novel mathematical formulations that are still the basis of ongoing research by academics across the world.

Energy Storage for Renewables: As part of the Energy Transformed Flagship Andreas worked in a multi-disciplinary team over several years to develop better models and algorithms for managing battery storage in conjunction with renewable energy sources. This work initially focussed on sizing storage, included developing a control algorithm for wind energy smoothing that has been implemented and tested at the Hampden wind farm, and has been expanded to dealing with solar-diesel remote area power systems with storage. The technology has been licensed to eCoult with restrictions places on the IP so that only limited publications have been possible.

Simulation: Andreas has been involved in a number of commercial projects involving the use of discrete event simulation across a range of application areas. This has included simulating the movements of inmates in a remand centre, modelling of a proposed new light rail (tram) line in Sydney, acting as expert witness in dispute arbitration about the performance Murrin-Murrin Nickel processing plant with respect to the simulation used to predict performance; and assessing simulation models built by a consultant on behalf of the company that had commissioned the simulation.

Scheduling At Trade Events: Tourism Australia organises very large trade events each year, involving buyers and sellers of tourism products. The smooth running of these events at which large amounts of business is transacted, requires the scheduling of many individual meetings based on the preferences of hundreds of buyers and sellers. CSIRO’s SATE software has been used for many years by to successfully schedule these events.