Hercules Routes: Presentation of business requirements, business processes and statistics

Abstract

The purpose of this document is to describe the business requirements identified during the development of the route planning and passenger transport framework system "Hercules Routes", identified as GINOP-2.1.7-15-2016-01364, and to describe the business processes and statistics of the system resulting from its implementation. The operation of the system was tested and statistically evaluated in 5 locations (industrial parks) over two years, between 01.01.2020 and 31.12.2021.

Introduction

Since the emergence of industrial parks in our country, intensive longdistance transport has been developed around industrial parks to cover the needs of commuting to and from work. These longdistance services are, with some exceptions, typically by bus and can cover an area of up to 50-60 km around the industrial estate. Unfortunately, the transport service providers contracting with the factories have no interest in optimising these routes, as they typically contract for fixedroute services, including the distance travelled. Such fixedroute services result in underutilised travel capacity and often half-empty services, imposing unjustified costs on the contracting factories and causing unjustified pollution and noise in the surrounding villages. To make matters worse, the factories located in industrial parks do not group their transport needs together but order them separately from the transport service providers, thus increasing travel costs and burdening the environment.

Focusing on this problem, our aim was to implement a service package that optimises longdistance bus services by aggregating the transport needs of industrial parks and then working with contracted transport service providers. Although readymade software systems for road freight transport have been available for a long time, at the beginning of our work in 2016 there was no solution available on the market for passenger transport, so we decided to develop an R&D-based system of our own design.

The framework we developed is based on artificial intelligence adapted to changing weather and traffic conditions, which is able to predict the average speed of the route network edges at any given time, and thus the expected arrival times, from continuously logged GPS data. We formulated the vehicle routing problem as R&D-based, multiple varieties of mixed integer linear programming model to achieve guaranteed quality results. The expected high computational performance was ensured by implementing our own R&D-based heuristics. The Hercules Routes framework is available as a cloud service through a browser and various applications for companies transporting their employees and for companies contracted to transport their employees. The system, running in the cloud, optimises the resources available (passenger buses) to transport workers to and from work in the most efficient way possible (number of people transported, km charge). By choosing the right vehicles and routes, significant cost savings can be achieved on the employer side, on the passenger transport service provider side, and in terms of emissions and noise pollution. Transported workers can see up-to-date information on their journeys via a mobile app, and drivers can receive information on the route and the people they are transporting via a driver's mobile or tablet app. All this is combined with a complex informatics service that ensures the organisation and punctuality of transport, the information of those involved in the transport of passengers, and the emergency alerts and warnings to be issued in the event of an incident.

Conclusions

In this document, we have reviewed the business requirements identified during the development of the route planning and passenger transport framework system "Hercules Flights", identified as GINOP-2.1.7-15-2016-01364, and the business processes and statistics of the resulting system.

Our key measurable expectations for the system were:

• Guaranteed high calculation accuracy: quality >= 99%, gap < 1%
• High performance computation, i.e. short computation time: t <= 90sec
• Expected 20% reduction in cost and environmental impact

These measurable results have been analysed on the basis of statistics from 5 live sites (industrial parks) for the last two years, i.e. from 01/01/2020 to 31/12/2021:

• Based on the measured results, we can state that both the accuracy of the calculation results and the calculation time have been met in all cases, and in the case of smaller industrial parks, shorter calculation times and often only 100% quality results were obtained.
• Our expectations for cost reductions and environmental impact reductions were significantly influenced by the transport needs of the factories and the route network in the vicinity of the park. In terms of cost savings, environmental pollution savings and noise savings per industrial park, we achieved the expected average savings of 20% in all cases, but for some parks we exceeded this level by far and achieved average savings of over 40%.

Overall, our work has been successful and we are confident that the framework will be used in other industrial parks too in the hope of a more economical, greener, flexible usage of route planning and passenger transport services. These principles are in line with the EU's environmental and sustainability ambitions.

The development of "Hercules Routes" was supported by the Government of Hungary and the European Union, with the help of the GINOP-2.1.7-15-2016-01364 grant.