Green Location Routing: Competitive Bi-objective Bi-level Model and Solution
This paper proposes a competitive bi-objective bi-level mixed-integer linear programming model for green facility location-routing problems between a manufacturer and a distributor, which is solved using a novel Bi-level Multi-objective Chemical Reaction Optimization algorithm that outperforms existing meta-heuristics on benchmark tests.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Every time a package travels from a factory to a doorstep, a complex chain of decisions determines its path. Someone must decide where to build the warehouses that hold the goods, and someone else must decide which trucks will carry those goods to customers and in what order. In the real world, these decisions are rarely made by a single person sitting in a quiet office. Instead, they are often the result of a tug-of-war between different companies. One company, usually the manufacturer, wants to set up facilities to move goods efficiently. Another company, the distributor, wants to pick up those goods and deliver them while making a profit. When these two groups act independently, their choices can clash, leading to higher costs and, increasingly, more pollution. As the world grapples with climate change, the transportation sector has come under intense scrutiny for its contribution to greenhouse gas emissions. The challenge for modern logistics is no longer just about moving things as fast or as cheaply as possible; it is about doing so while minimizing the damage to the atmosphere. This requires a delicate balance between economic goals and environmental responsibility, a balance that becomes even harder to strike when two competing businesses are involved.
Researchers at the University of Kurdistan have tackled this specific tension by creating a new way to model these interactions. They focused on a scenario where a manufacturer acts as the "leader" and a distributor acts as the "follower." The manufacturer decides first, choosing where to open new depots and how much product to send to them. Once those decisions are made, the distributor looks at the available options and decides how to route their vehicles to serve customers, aiming to maximize their own profit while keeping their own costs and emissions low. The researchers built a mathematical framework to simulate this two-step process, but with a twist: they treated both companies as having two goals at once. The manufacturer wants to keep their fixed costs and carbon footprint low, while the distributor wants to maximize profit and simultaneously minimize the emissions from their trucks. This dual focus on money and the environment creates a complex puzzle where improving one goal often makes the other worse.
To solve this puzzle, the team developed a new computer algorithm inspired by the way chemical reactions work. In nature, molecules collide, break apart, and combine to form new substances, constantly seeking a state of lower energy and greater stability. The researchers translated this biological principle into a digital tool they call a bi-level multi-objective chemical reaction optimization. In their system, potential solutions to the logistics problem are treated like molecules. These digital molecules interact with each other, merging, splitting, and shifting their properties to explore millions of possible combinations of depot locations and truck routes. The algorithm is designed to find the "sweet spots" where both the manufacturer and the distributor can achieve their goals as well as possible without destroying the other's interests. It does not just look for one perfect answer, but rather a range of balanced options that show the best possible trade-offs between cost and carbon emissions.
The team tested their new method against two other well-known computer techniques used for similar problems: a genetic algorithm that mimics evolution and a search method called tabu search. They ran simulations using various sizes of logistics networks, ranging from small setups with a few factories and customers to large, complex systems with many more. The results showed that their chemical reaction-inspired approach was significantly faster than the others. In tests involving different numbers of locations and vehicles, the new algorithm consistently found high-quality solutions in less time. For example, in one test with a moderate-sized network, the new method found a solution in just over two-tenths of a second, while the other methods took longer. More importantly, the solutions it found were not just fast; they were effective. The algorithm successfully identified routes and depot locations that kept costs down and emissions low for both companies, proving that the two-step decision process could be optimized without sacrificing one goal for the other.
The researchers also examined how sensitive their model was to changes in the real world. They adjusted variables such as the cost of transporting goods and the environmental impact of opening a new depot to see how the system would react. They found that the algorithm remained robust even when these costs fluctuated wildly. When transportation costs rose, the system naturally adjusted the routes and depot choices to compensate, keeping the overall plan feasible. This suggests that the method is flexible enough to handle the unpredictable nature of real-world logistics, where fuel prices and environmental regulations can change overnight. The study confirms that by modeling the relationship between a manufacturer and a distributor as a cooperative yet competitive game, it is possible to find strategies that benefit both parties and the planet.
While the study offers a powerful new tool for planning, the authors are careful to note its limitations. Their model assumes that both companies share information perfectly and act in a way that is mutually beneficial, a scenario that might not always happen in a cutthroat market where one company might hide its true costs or act aggressively. They also focused on a single distributor, whereas real markets often have multiple competing delivery companies. Despite these simplifications, the work provides a clear path forward. It demonstrates that the complex, hierarchical nature of supply chains can be understood and optimized using advanced computational methods that respect both the bottom line and the environment. By treating the logistics network as a living system of interacting parts, the researchers have shown that it is possible to design supply chains that are not only efficient but also sustainable, offering a blueprint for a greener future in global trade.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.