According to the 2024 Círculo Logístico Barometer conducted by SIL 2024, 89.5% of companies believe that innovation has been key to improving logistics efficiency, with automation and robotics playing a major role. Additionally, 49.2% of respondents indicated that Logistics 4.0 has had a significant impact on supply chain performance.
The adoption of advanced technologies like Decision AI—a discipline that combines various AI techniques to support decision-making—is transforming logistics and boosting business profitability. It directly addresses the critical question: “What should we do?” to optimize processes and reduce costs.
What do we mean by a transportation problem in Logistics?
Transportation logistics is typically divided into two categories: primary transportation and secondary transportation.
Primary transportation refers to long-distance movement between suppliers and warehouses, or between central warehouses and regional distribution centers. Common challenges in primary logistics involve long distances, planning complexity, limited numbers of origins and destinations, and the large capacity of transportation modes.
The key challenge is to strike a balance between meeting delivery deadlines, managing transport costs, fleet size, and the number of transfers or transshipments required.
Secondary transportation, or last-mile delivery, involves moving goods from a local center to the final customer. Depending on the business, this could mean supplying retail stores from a regional warehouse or delivering parcels to customers within a city. This type of transportation is characterized by short distances, detailed scheduling, a limited number of origins but a large number of destinations, and smaller-capacity vehicles.
The challenge here lies in designing efficient delivery routes that allow a single vehicle to serve multiple customers. These routes must balance delivery time windows, vehicle capacity, total distance traveled, and overall delivery time.
There are also other types of transportation problems that don’t fall neatly into these two categories. For instance, goods may be transferred between warehouses or stores, or return trips may be used for secondary tasks like collecting empty pallets or reusable packaging.
How Does It Work?
The logistics transportation problem is one of the most studied issues within Decision AI. Commonly, this problem is referred to as the vehicle routing problem (VRP), and it has many variations:
There are also various variations within each of these categories. For example, a VRPPD may include a restriction where the last item loaded must always be the first one to be unloaded. This restriction effectively represents the loading and unloading of pallets on a truck.
The different variations of the problem can also be combined to create much more complex theoretical problems.
To tackle these types of problems, theory proposes a significant number of techniques and methods. The simplest and most direct approach is to study all possible solutions. For instance, to create a route visiting three clients—A, B, and C—we can analyze the six possible solutions (ABC, ACB, BAC, BCA, CAB, CBA) and choose the best. As the number of possible routes increases, the total number of solutions to analyze grows exponentially.
This highlights the clear need to explore new strategies for finding solutions more efficiently, including:
Generating solutions randomly and selecting the best one
Applying a sensible criterion to build the solution (e.g., always going to the closest customer left to visit)
Making small modifications to good solutions in an attempt to achieve better results
Identifying routes in good solutions and combining them to obtain better solutions
Identifying routes in poor solutions to avoid using them
Based on these ideas, a range of methods or heuristics with varying complexity have been developed: random walk, greedy algorithm, simulated annealing, ant colony optimization, and genetic algorithms.
Another option for solving these types of problems involves exact methods, where the problem is defined as a series of equations to be solved. Techniques such as Constraint Programming and Mixed Integer Linear Programming aim to find the best solution that satisfies a set of constraints
Is theory enough to solve my real problem?
The theoretical problems studied in literature are never exactly the same as the real-world problems faced by companies. Depending on the type of business and their commitment to customers, a company’s actual logistics problem almost always includes specific rules or peculiarities.
A real-world problem can involve a wide variety of vehicles, such as trucks, buses, ships, planes, trains, or even pipelines and oil pipelines. Drivers may also be subject to specific rules, such as working hours, overnight costs, daily or weekly driving time limits, and certifications for certain types of cargo.
For this reason, theoretical methods always need to be adapted to fit the real cases that companies face. By relying on theory and previous experiences, the most suitable methods for problems similar to the one being solved are often identified, but they must be adapted to the specificities of the real problem in order to translate into a competitive advantage for businesses.
What Impact Does This Have on the Business?
The advantages and impact of Decision AI on the logistics sector are crucial.
- By finding the optimal solution to my transportation problem, it may result in traveling fewer kilometers, which translates to reduced fuel consumption and, therefore, lower transportation costs.
- Another key indicator impacted by this technology is the reduction of immobilized assets and cost savings. By finding the most efficient way to transport goods, the solution may involve a reduction in fleet size.
- Furthermore, by optimizing this entire process, we can deliver more packages and do so more efficiently, leading to an increase in productivity.
These are just a few of the key benefits for companies in the logistics sector, but the possibilities are vast.
How to Carry Out a Project to Solve These Problems?
To apply and solve these real logistics problems using Decision AI, transportation logistics projects are approached in several sequential stages.
If you believe this technology could be applied to your specific problem or simply want to learn more about how Decision AI is already impacting business processes, you can schedule a smart session for you and your team with our experts in this technology.
