00:29 Cédric, perhaps you can start by telling us a little more about Kardinal?
02:49 Using the human brain as an addition to an optimization, is that something you would agree on?
03:29 Why is real time route optimization interesting to you from a supply chain perspective?
05:13 How has route optimization developed over the last few years?
08:07 Which are the key players that have driven this growth in route optimization?
12:45 From a technical perspective, what sort of challenges does work in real time introduce?
15:31 Users can nowadays control pieces of data. Do you think this has a positive impact?
19:52 From a data perspective would you say that we are now too reliant on too few companies like Google, Amazon or Microsoft?
22:06 In terms of R&D, what do you see as being the biggest areas of interest over the next few years?
With the latest advances in crowdsourced data, we are now able to forecast the impact of traffic congestion and optimize our routes more accurately than ever before. For this episode of LokadTV, we are joined by Cédric Hervet to discuss how real-time route optimization has changed the way delivery companies operate.
Cédric is the Co-founder and Head of R&D at Kardinal, a route optimization software company. Kardinal’s philosophy is that humans shouldn’t be removed from the equation when it comes to optimizing routes and delivery times, as the human brain is equipped to deal with situations and has prior knowledge that an algorithm simply can’t replicate. Instead, Kardinal seeks to combine cutting-edge technology with human intelligence for peak optimization.
From a supply chain perspective the time-scale isn’t the same, as route decisions can be re-challenged every minute. Lokad optimizes decisions for the next day or even for the next year ahead, for example rebalancing inventory between locations such as stores and warehouses, which naturally depends heavily on the crucial agility and flexibility that route optimization softwares can provide. More agility for supply routes means a greater range of possible optimization at a lower cost.
A major problem is the number of stops that need to be visited by a transporter. There are often such an overwhelming number of combinations that it would take years to compute, requiring algorithms that work more intelligently than just numerating. Accurate data is key here, otherwise the probability of unfeasible routes being suggested is very high.
In addition, we discuss how Google has been a major player in this domain, absolutely transforming how routes are optimized and inspiring many other companies. We also talk about the various complexities that come with working in real-time, whether today we rely too much on the major players, like Google and Amazon, and whether map data should be a “common good”.
To conclude, we go into more detail about the exciting new developments in AI, such as reinforcement learning and the “quantum supremacy” announced by Google and how they can be applied to the various problematics that remain in route optimization.