Created 26 world records | Rookie R&D Global 60 years of leading technology in vehicle routing planning

The industry believes that the IoT technology products such as the Rookie Future Park and the “Logistics Eyes” are accelerating the digitization of logistics elements, and the artificial intelligence technology will be the key to realize the intelligent dispatch of logistics elements. The rookie algorithm has set a world record, showing the deep accumulation of rookie in artificial intelligence technology.

Rookie creates 26 world records

According to reports, the breakthrough of this problem is the rookie artificial intelligence department warehouse intelligence and vehicle path planning algorithm team. This team specializes in artificial intelligence technology, and continuously explores the effective integration of optimized search and machine learning technologies, seeking optimal vehicle path optimization solutions to improve logistics distribution efficiency and reduce costs.

Team leader Hu Haoyuan introduced that the rookie vehicle path planning algorithm has been applied to many businesses. In the vehicle distribution process, the vehicle path planning algorithm can effectively reduce the number of vehicles used and the distance traveled by the vehicle; in the internal picking operation of the warehouse, the vehicle path planning algorithm can reduce the walking distance of the picking personnel. In addition, the vehicle path planning algorithm can also help takeaway distributors to plan delivery routes, thereby improving customer experience and significantly reducing distribution costs.

Vehicle routing problem is one of the most classic problems in the field of operational optimization. Hu Haoyuan said that a logistics company needs to distribute to 1000 outlets as an example. Its distribution route is countless. If you can find the most efficient distribution route, it will greatly improve logistics efficiency and reduce costs.

Hu Haoyuan said that if the global computing of 5 billion computers is used to calculate 2 billion times per second per computer, verifying the efficiency of each distribution path one by one requires at least trillions of years, which is a task that is impossible to accomplish.

Public information shows that at this stage, researchers in various countries are using algorithms to find the best distribution path. The European independent research institute SINTEF initiated and managed the vehicle path planning algorithm evaluation platform in 1999 and is considered to be the most authoritative evaluation platform in the world. Global researchers can download data sets to impact this review list.

As of October 9, 2018 Beijing time, the rookie won the championship in the 26 data sets of the evaluation list. This means that in 26 logistics scenarios, the rookie algorithm can use the fewest vehicles, the shortest mileage, and complete the delivery task.

City distribution order delivery cost reduced by 10.3%

Public information shows that the vehicle routing problem was raised by two Western scholars in 1959. Over the past sixty years, Western research institutions have accumulated and accumulated a great deal of experience.

In the nearly two decades since the establishment of this list, the world's top research institutions, well-known scholars, and established logistics companies such as the United Kingdom and Poland have continuously set records and broke records. They have long occupied the top spot, and none of the Chinese researchers have been on the list.

Speaking of this list, Hu Haoyuan said that although the global computer computing power and algorithms are improving, the experience of sedimentation and accumulation in Western countries is rich. Some Western researchers have a national background and will use super-calculation centers. Researchers, this leaves very limited space for ordinary Chinese researchers.

Hu Haoyuan said that a rookie had tied many world records on this list a year ago, but has not been able to break through for more than half a year.

"We don't need to brush the list, but we want to use this data set to verify the reliability of our algorithm, and after verification, apply it to the actual business." Hu Haoyuan said that this means that the rookie must use a normal server, not for Brush the list and use a supercomputer.

According to Hu Haoyuan's introduction, the rookie balances the breadth and depth of the search based on the probability of prediction by integrating machine learning ideas into the control of the search strategy, thereby greatly reducing the space required for search while retaining the ability to jump out of local optimum. In the end, many world records were created.

“China has the world's largest logistics market, involving a large number of vehicles and personnel. The selection of routing routes is optimized. With the arrival of the new logistics era, the demand for instant delivery is strong, which also involves the optimization of the distribution path of a large number of foreigners. A breakthrough in path planning will bring about a very large social effect.” Hu Haoyuan said that by applying the vehicle path planning algorithm in the retail through-city distribution business, the order delivery cost has been reduced by 10.3% and the warehouse is promoted. The efficiency of cargo circulation has increased, and the warehouse collection turnaround time has decreased by 57%.

It is reported that the rookie launched the IoT strategy this year, and has released technical products such as the future park and logistics eye, and continuously digitized the logistics elements. After completing the digitization of logistics elements, how to use artificial intelligence technology for scheduling is a key issue. The rookie is also using IoT and artificial intelligence technology to promote the industry's continuous intelligence, leading the industry to build a smart logistics network.