Without big data and artificial intelligence, you may not be able to eat if you are hungry

【51CTO.com original article】 Friends born in the 1980s or older must have watched the classic skit called "Working Adventure" in the 1996 Spring Festival Gala. At the end of the skit, actress Zhao Lirong held a pot and a pan and said the line "Do you still need to call a taxi to buy tofu?!" 20 years later, you don't need to call a taxi to buy tofu, and someone will deliver it to your home. The fast and convenient food delivery industry has been born, developed, and grown in the Internet era. In the past period of time, the food delivery industry was mainly driven and developed by local promotion and execution. In the future, the second half of the food delivery industry will not rely on execution, but on technological innovations such as artificial intelligence and data analysis.

Recently, 51CTO had the honor to interview Mr. Zhang Hao, Vice President of Technology of "Ele.me", and listened to his views on the combination of the food delivery industry and artificial intelligence technology, and the new service and technology standards of the food delivery industry. Without artificial intelligence, food delivery meals will not be delivered on time, and you may not be able to eat when you are hungry.

[[196144]]
Zhang Hao, Vice President of Technology at Ele.me

【Zhang Hao, guest speaker at WOTI2017 Global Innovation Technology Summit, Vice President of Technology of "Ele.me", responsible for the company's artificial intelligence and big data construction. He led the team to apply machine learning in logistics scheduling, pressure balance, recommendation search and other scenarios, established a complete data operation system through data mining, improved operational efficiency, and used data and intelligence to drive business development. He has more than ten years of practical experience in machine learning, data mining, and distributed computing. He has served as Senior Director of Didi Research Institute, Senior Data Scientist of Uber Big Data Department and LinkedIn Search and Analysis Department, and Senior Data Engineer of Microsoft Speech Recognition Group. 】

Many years ago, most restaurants received takeout orders by phone, which was inefficient and often busy during peak hours. "Ele.me" saw this opportunity and wanted to try to provide merchants with a simple system and an online platform so that they could receive orders online, thus solving problems such as busy lines and order printing. In April 2009, the "Ele.me" online ordering platform was established in Shanghai.

With the popularization of mobile Internet, the supply and demand of online food ordering are booming, and food delivery boys are all over the country. In the eyes of people who are not familiar with the food delivery industry, food delivery and AI are two terms that are completely unrelated. The basis of AI is big data, and AI can shine wherever there is massive data. Ele.me's AI applications are mainly reflected in three directions: transaction efficiency, operational efficiency, and logistics efficiency. The essence of food delivery is the combination of transaction and logistics. The offline part is very heavy, reflecting the management and operational capabilities; and AI can increase the efficiency of offline operations by multiples by increasing data, and in some aspects it will completely replace offline work, such as intelligent scheduling.

For human dispatchers, the peak hours at noon and in the evening are huge challenges. For example, during the peak meal time, a dispatcher has to dispatch one order every 6 seconds. He needs to consider the number of orders the rider has, the familiarity of the route, etc. This job will undoubtedly make humans overwhelmed. But for the artificial intelligence ET dispatching engine, handling these problems is easy.

How does ET achieve intelligent dispatching and ensure the highest efficiency? Simply put, ET will insert new orders from the distribution station into the existing tasks of each rider, re-plan a round of the shortest delivery route, and compare which rider has the shortest additional time. In order to accurately estimate the additional time, ET needs to know the delivery speed of 1 million restaurants across the country, the riding speed of more than 1.8 million riders, and the time it takes for each customer to go downstairs to pick up the meal. Generally speaking, the waiting time for restaurant meals accounts for one-third of the entire delivery time. If ET wants to improve the efficiency of riders, it must accurately estimate the delivery time to reduce the waiting time of riders, but it cannot let the meals wait for people, otherwise the meals will get cold. In order to calculate the delivery time of riders, ET also needs to know the delivery speed of each rider in different areas and different weather conditions. But when the meal is delivered, the customer may not come to pick it up immediately. The customer may have to wait for three elevators to come down. All these ETs need to be calculated.

With the help of the artificial intelligence ET scheduling engine, Ele.me has put forward the vision of "Make Everything 30'". Artificial intelligence plays a key role in intelligence and efficiency. In the short and medium term, in the local life scenario, it allows users to find their favorite food and daily necessities more effectively and quickly, and have them delivered to them within 30 minutes. In the medium and long term, when technology and regulations and laws are in place at the same time, robots and drones will gradually become popular, completely subverting the labor-intensive production mode of logistics.

The entry of artificial intelligence into the food delivery industry is not a one-off event. Logistics is an industry with thousands of years of history. The improvement in efficiency after the intervention of AI is not a 100-meter sprint, but an accelerated race. Once you get ahead, manual scheduling will not be able to catch up. In just 30 minutes, restaurants, riders, shopping malls, customers, traffic jams, roadblocks, elevators and even the doorman downstairs will have unpredictable effects on service quality. Using data mining, machine learning and operations research technology, combined with offline operations, intelligent scheduling can maximize the accuracy and rationality of scheduling and continuously improve service efficiency.

On July 21-22, 2017, the WOTI Global Innovation Technology Summit hosted by 51CTO will be held at the Beijing Renaissance Hotel. The summit will focus on three major themes: machine learning, human-computer interaction, and intelligence + application. Dozens of expert guests will bring many exciting technical content sharing. At that time, Mr. Zhang Hao will share the keynote speech "Takeaway Deliveryman Sent by AI" with guests at the main venue of the *** Forum, and elaborate on how artificial intelligence can improve takeaway delivery services and set new industry standards. 51CTO sincerely invites you to attend the conference and share the joy brought by technology with us.

[51CTO original article, please indicate the original author and source as 51CTO.com when reprinting on partner sites]