Brazil-based CFlex is now Rail Movement Planner. Rail Express finds out what prompted the change as the company continues its international growth.
Brazilian technology company Rail Movement Planner (RMP) has been in the rail game for a while now, but not as you’d know it. The company was founded as CFlex MPC in Campinas, Brazil (around an hour’s drive north of Sāo Paulo) in 2015, where it quickly made a name for itself for its CFlex Movement Planner product.
The company has now changed its name to Rail Movement Planner, having rewritten and evolved the CFlex Movement Planner consistently since 2015.
“The name ‘CFlex’ was too generic,” explains RMP president Elesbao Oliveira. “If you Google it you can find many companies that are not related to the railroad world.
“So, to improve our presence in the market and increase our visibility, the Rail Movement Planner company was born on June 3, 2019.”
Rail Movement Planner (RMP) is a state-ofthe-art solution for real-time circulation planning and train dispatching that can provide fully integrated timetabling, conflict detection and problem-solving services for rail operators.
This helps to improve the visibility of planned train circulation, allowing train controllers to increase the average speed of trains, which can lead to several productivity benefits.
It’s also possible to create and configure a regularity operational scenario where the RMP engine will prioritise the regularity of the train circulation.
“RMP optimises train circulation automatically or at the train controller’s request, delivering optimised and feasible train circulation plans in just a few seconds,” Oliveira says. “These plans take into consideration all constraints imposed by the dynamics of any complex train operation and fully comply with customer quality standards and safety rules.”
The system is already in place at several railroad operating centres worldwide, including Australia, where the solution is used by mining powerhouse Rio Tinto to provide effective planning of the company’s freight trains.
In the railroad’s operation centre, train controllers use our solution to plan the train circulation and have a great visibility of what’s coming in the next hours or days,” Oliveira says. “It’s also possible to apply restrictions to the plan — track blocks, speed restrictions, mandatory movements, high priority trains, trains’ dependencies — to simulate and validate the planned circulation.”
The company also operates internationally in territories such as Argentina, Chile and Canada, and has received significant government support in the domestic Brazilian market from the São Paulo Research Foundation (FAPESP, Fundação de Amparo à Pesquisa de Estad de São Paulo) and Studies and Projects Financing Agency (FINEP, Financiadora de Estudos e Projetos).
The collaboration took place over three phases: concept, development and marketing, resulting in the successful rollout of the RMP product.
“Our partnerships with FAPESP and FINEP have served to successfully evolve and commercialise Rail Movement Planner,” says Carlos Carneiro, RMP vice president. “We also have another important partnership with the State University of Campinas, one of the best Brazilian universities. This partnership allows RMP access to brilliant human resources and cutting-edge researchers.”
RMP uses three levels of optimisation that are suited to different levels of train planning.
The first level of optimisation is the Basic Engine, which is built to address around 95 per cent of railroad planning issues, according to RMP, taking factors such as network topology, train specifics and basic operating rules into account.
“Basic Engine uses a meta-heuristic algorithm that is very effective and fast, with a typical response time of around five to 10 seconds” Daniel Oshiro, RMP project manager explains.
Adaptive Engine, the second optimisation level, is purely heuristicbased design capable of accommodating algorithms written in Java or Drools. It is intended to solve issues that cannot be properly solved by Basic Engine. Examples of this can include issues introduced through the specific idiosyncrasies and preferences of train controllers, or other special operational rules not otherwise covered by Basic Engine implementation.
The third level of optimisation refers to RMP’s Meta Planning Engine (MPE). Indeed, MPE is a powerful software optimisation architecture that allows RMP to run in parallel to several instances of the Basic Engine to address complex network issues. MPE also encompasses an Artificial Intelligence (AI) evaluator to select the optimal plan for users.
“The RMP integration layer encompasses mechanisms to provide and consume services, message mediation, service orchestration, service governance, business process management and service monitoring across all three optimisation levels,” Oshiro says. “RMP is scalable, expandable and was designed to fit any size of railroad operation.”
It also includes support for exploratory data analysis (EDA), model-driven architecture (MDA), and other enterprise integration patterns. It also provides support for industry standards such as WS-*, representational state transfer (REST), and other binary and nonbinary protocols. Quality of Service (QoS) capabilities such as security, reliable messaging and throttling are also built into RMP.
To ensure RMP runs as effectively as possible, round-the-clock services support and comprehensive user training are also available. Training is provided for managerial groups, train control groups, IT and other key users.
The managerial training is attended by managers and operation coordinators and focuses on introducing system support philosophies, basic architecture and functionalities to management.
Training for train control groups discusses system usability and how to generate effective train circulation plans. IT and key user training is an advanced course that focuses on high-level system functionalities, database maintenance and troubleshooting techniques for end users.
Overall, RMP is confident that it will continue to make a name for itself in the industry, even if the name itself has changed.