We have created a challenge named Race Against the AI. The goal of the challenge is to create an optimal shopping plan for purchasing Christmas presents. The challenge is to minimize expenses while still having time to purchase each specified gift in time.
We gave this challenge to randomly selected f2f-event participants. They raced against an optimization algorithm. With an analytical approach and a bit of luck, the top 13% of them found the optimal shopping plan. This took them 3 minutes on average. The algorithm, on the other hand, spent less than 10 seconds to find the optimal plan and to prove that there is no better option.
The idea was to showcase concretely the benefits of optimization algorithms and it surely worked. When the participants had raced against the AI it was clear to them that optimization algorithms can tackle challenges like this better and faster than they could.
Optimization algorithms can be used to improve planning quality as well as saving time for the human expert. The real-world problems are much more complex than the problem used in this challenge. The number of possible solutions grows exponentially with the size of the problem, but a good optimization algorithm can find the best solution very quickly despite this.
In our challenge the planning is done in IBM Planning Analytics and IBM Decision Optimization. The optimization algorithm can be run by pressing a button in Planning Analytics. Using this setup, a human expert could edit values, save scenarios, create visualizations, and quickly obtain the optimized solutions for each scenario.
On the video you can watch what the challenge looks like and how the algorithm provides the optimized solution to a new scenario.
If you are interested in optimization and Planning Analytics, please don’t hesitate to contact us.