In a previous blog, we addressed the general and specific needs of businesses related to transportation optimization. It already pointed out the fact that one optimization cannot solve all the problems of transport optimization. That is, mathematical modeling and the optimization algorithm really matter, not to mention if you use optimization with a small ‘o’ or a big ‘O’?
Difficulty of selection between optimizations
So how should I choose an optimization that gives me a superior competitive advantage or customer experience over my competitors? Should I follow the mainstream and use the same software as the others, or should I do something groundbreaking and build my own?
There is a lot of off-the-shelf software on the market that includes optimization functionalities. You can evaluate these and compare the fit to your business with your team for example, or you can take an impartial consultant to make a comparison for you.
However, keep in mind that your business needs will drive the implementation of your software and optimization - not the other way around, (why try and drive a square peg through a round hole) you run the risk of ending up in a non-optimal situation for your business. Or a situation where you are copying the same solution your competitors already have which may not be fit for your purpose.
A customized optimization solution is worth of considering:
- if you really want to have a solution that is relevant and suitable for your business. A solution that is also easy to deploy, with terminology and usability tailored to your company. In this context, it is also good to consider the costs of deployment and training of staff.
- if transportation optimization is combined with other functions to be optimized or balanced, such as further processing of the transported goods. There may be two or more optimization problems that need to be combined so that both or all achieve the best possible result for the business.
- if you want a solution that other players in the industry do not have. Optimization is a source of competitive advantage. Today’s advanced optimization solutions help companies identify expansion opportunities, explore collaborative partnerships, and investigate new business models. They are a platform for innovation, growth, and long-term success.
In the optimization solution, the business rules are transformed into mathematics. As long as the same rules apply, the model will cope very well with new, unprecedented situations (such as the corona pandemic). Here, optimization differs from machine learning, where it is not usually possible to react correctly to phenomena outside the historical data.
If you want to learn more about the business value of optimization technology, I can recommend our partner, Gurobi's online webinar (about 50min). This webinar is about the applicability of optimization solutions to different problems and it opens up direct and indirect benefits to the business.
What can you expect from a next generation optimization solution?
There is a lot of off-the-shelf software available for route and delivery optimization. When these softwares runs out of ability to meet your needs, or other transportation optimization types of activities are at the core of the business, it is appropriate to consider optimization with special features. Ready-made algorithms can be applied to these situations as well, you just have to know how to apply them.
Transportation optimization tailored to your business can solve a wide variety of special cases in addition to the traditional ‘Travelling salesman’ optimization problem, the only limitation is imagination. Here are some examples of the special needs we have encountered over the years:
- Balancing the number of deliveries between different weeks and days so that the use of available working time becomes more efficient and balanced - reducing overtime.
- Optimization of the container export-emptying-washing chain and selection of the best emptying location from several options, taking into account the transition and reception costs.
- Ordered taxi rides, so-called trip merge taking into account routes, schedules, waiting times, allergies, etc.
- Delivery of packages taking into account strict delivery time windows and same day orders
- Choosing the optimal car for the delivery when there are dozens of cars on the ground and orders come in every minute.
- Very versatile capacity and compatibility constraints , such as the choice of fleet to be used based on the accessories available in the vehicles and needed at the site.
Metaheuristics and artificial intelligence
Since a single optimization algorithm cannot guarantee a solution to all optimization problems, an optimization expert must be able to apply multiple algorithms. Or an expert needs to develop metaheuristics, as we have developed in the field of transportation optimization. In this way, we have created a sophisticated transportation optimization entity that can be fed with both complex and versatile optimization problems.
Where heuristics means a simple rule of thumb, metaheuristics means a rule of thumb about which rules of thumb should be used to find the optimal solution. The goal of metaheuristics is to ensure that the algorithm does not get “stuck” in the environment of a good solution, but looks for even better solutions a little further away, yet using computational power sensibly.
Let's play with this a bit and expand somewhat:
Imagine a mountaineer looking for the highest mountain possible without a map. It would be impossible to study every square kilometer of the earth, on the other hand, if the search is limited to the territory of Southern Finland, the result will be modest. Metaheuristics are used to select a random number of promising square kilometers and limit the search to their immediate surroundings, always searching for the highest mountain in sight until one is no longer found.
Is metaheuristics and optimization about AI, artificial intelligence? Here, I would like to hand over the responsibility for deciding this, to you, the reader. From an optimization expert perspective, optimization is an old technology, we’ve been doing optimization for decades. But as often in the IT industry, there will be various megatrends, such as AI. However, the term has been coined much later than the origins of optimization, so we will leave it to you to decide whether or not optimization is an AI technology, according to how you define artificial intelligence.
All in all, if a company chooses the wrong optimization solution, it can end up in a situation where the last percentage improvements are not received. What could be achieved with them? More customers, service extensions or better customer service? So much more could be achieved!
Could we boost your business through optimization?
Read more about our optimization current state assessment.