Advanced Data Analytics for Transport Route Optimization

May 5, 2020 By Natalia Andueza

One of the main challenges faced by any company that manages transport is route optimization. It is necessary to find the optimal balance between the service provided to customers and the costs involved in delivering it, or in other words, the main goal is to achieve the maximum possible profitability in freight transactions.

To achieve this optimization of profits, you must start by finding the best combination of resources and loads, taking into account all the external conditions and restrictions surrounding this process. This task requires extensive knowledge of the sector, the market, the available resources, the types of loads, and our customers’ needs, among many other factors, since each piece of the puzzle affects how the next one fits.

Analyzing all these variables is often a tedious process for traffic managers, who often rely on market optimization tools that fall short when it comes to applying the right combinations of information to make the most of the service. There are many variables and data points to take into account, which means that transport route optimization tools must, at a minimum, offer:

Integration with various data sources, to have access to all the necessary information
Data processing capability in both batch and streaming modes, in a constantly changing and moving environment such as transport
Powerful and intuitive visualizations for the user
Agile, user-friendly features for efficient use

What are the phases involved in a transport route optimization project?

LISDATA’s experience developing projects to improve transport processes has given us insight into the complex nature of the sector, allowing us to establish five phases for processing and analyzing the large volumes of data these processes generate:

Phase 1: it is necessary to carry out an analysis of the current situation, in order to gain a complete view of the business, along with all the conditions and restrictions that may affect it.
Phase 2: study of the data available to the company. At this point, the variables to include and the sources from which to obtain the information must be defined.
Phase 3: modeling of all the data, defining the infrastructure and architecture required for its development, and the features to be covered in the automation process. This is where artificial intelligence is incorporated into the model.
Phase 4: development of the visualization environment, either using the client’s own tools or through comprehensive developments by the LISDATA team.
Phase 5: implementation and validation of the solution at destination.

What technical implications does a route optimization project involve?

As we can see, achieving the ultimate goal of optimizing transport channels is a long road that necessarily requires a high level of technical expertise to tackle the project and the defined developments, as well as extensive knowledge of the sector and the business in order to link analysis with results. Within this framework, there are particularly complex and sensitive aspects that must be defined by joint multidisciplinary client-LIS teams, which can be summarized as follows:

Analysis of all internal data sources: tables, entities, data volumes, correlations, etc.
Analysis of the client’s business restrictions and constraints: key to effective optimization and planning.
Knowledge of the regulations affecting road transport at the European level, in order to take into account ‘external’ restrictions and constraints.
Study and analysis of external data sources to be incorporated, along with the technical complexity this entails, especially in integrating and capturing such data: traffic, weather, load boards, etc.
Study, definition, and development of algorithms that deliver optimal results to support the traffic manager, taking into account all the technical, sector, and business considerations mentioned above.
Defining, together with the client, the final visualizations or features of the project, provided these are not implemented on the client’s own tools (BI tools, TMS, market planners, proprietary tools, etc.).

With all of this, the client is able to implement a route optimization project, which, as mentioned earlier, allows them to reduce operating costs while improving their level of customer service.

What do advanced analytics and artificial intelligence bring to the table?

Once the data extracted from the company, sector, environment, socio-political climate, external contingencies, and other predefined variables have been analyzed, improvements can be achieved from different angles of the client’s operations, as varied as the services or transport verticals that exist. In models where the flow is more standardized, with pickups, sorting, and onward shipping from platforms to final delivery points, the improvement will be seen at the start of the operation and can be optimized periodically by applying learning techniques to the developed tool, enabling it to offer improvement proposals.

In models with daily scheduling and ‘cut-off times’ for receiving orders, generic planning will be enhanced and improved with all the variables to be incorporated into the model. It could even be possible to automate the process, so that once the scheduled time arrives, the system launches automatically and the results are presented to traffic managers for review, modification, and approval.

On the other hand, if the planning or assignment model is continuous and random, the system should always be on the lookout for new loads, continuously optimizing assignments in order to put forward specific proposals for each load received. In this case, the manager will receive proposals on their dashboard, which they can accept or reassign to better options, improving profitability and the service provided on the routes. This model requires more complex learning and intelligence, developed over a longer period of time and on an ongoing basis, but it will deliver excellent results in terms of cost improvement.

Conclusions

It is clear that the process of optimizing and planning routes in a transport management company is critical and strategic, which means that improvement projects in this area are equally critical and, therefore, require a high level of technical and sector/business expertise, which we achieve through client-LISDATA partnerships.

This thorough analysis generates a huge amount of information to process, resulting from the large volume of data generated, which is extremely important in an environment where both loads and vehicles are constantly on the move. In the transport sector, traceability is essential, requiring real-time data processing, accurate estimated arrival time calculations, and even on-the-fly changes to previous plans. This is where Big Data becomes essential.

Furthermore, by equipping the system that supports traffic managers with intelligence, it becomes critical to understand the restrictions and constraints that apply to the decision-making algorithm and the learning model chosen for implementation.

Likewise, defining the architecture and infrastructure is complex and must take into account the scalability and flexibility required in the future. In the same vein, the ability to deliver comprehensive developments or integrate with client tools makes the choice of technology partner for this type of project equally strategic.