From Digital city to Smart city

Wen/Li Wei

Academician of CAS, State key Laboratory of Software development environment, master Ren

Over the past 10 years, "Digital City", "Big Data City", "Intelligent city" and "smart city" have become the hot topics of academia, various pioneer forums and social media. In general, an industry or a region of the information process can be divided into two stages, the first stage is digital, also known as the Digital City (hereinafter referred to as DCity, which is the data cities), it includes the information sensing, storage, computing, transmission and control of the network and digitization, the second stage is intelligent, Is the higher stage of digitalization, also known as Smart City (hereinafter referred to as icity, that is, intelligent cities).

From the perspective of intelligent transformation of Digital City, the following paper explores several services that smart cities must provide, as well as the measures and methods that need to be taken to realize these services in the implementation of software engineering.

depth of Inquiry

Let's start with an example of what a deep query is and then see how to develop and use that query. If we input to Google: "I often nausea, headache, sometimes vomiting, see things unclear, occasional tinnitus, what should I do?" Google put "headache, nausea, vomiting," and so on as the key words, find out about 42,900 pages appear in the above keywords. While we marvel at Google's efficiency, we cannot call it "intelligent service" because of the problem we are most concerned about: what is the "I" Disease? How should it be treated? Did not get a straightforward and constructive answer like a friend, let alone a doctor.

Below, let's imitate the patient to go to the hospital procedure, design a simple "deep query" example. This query process consists of the following 4 steps:

newly diagnosed pilot: patients enter symptoms or feelings into the icity: "headache, nausea, vomiting, tinnitus and how to do", I hope that ici T y by analyzing these feelings and symptoms, to provide users with the equivalent of the initial diagnosis of the disease of the first diagnostic and recommendations (see figure). Icity should provide the patient with the main name of the disease, namely "brain tumors, Alzheimer's disease, cervical spondylosis or gastrointestinal cold." Patients choose one of 4 diseases, such as "brain tumors", and enter them into icity, which is called Association selection or association inference, because Icity Associates 4 different "entities" with common pathologies.

proposal: It consists of two parts, that is, to do the association first, then do logical reasoning. First, after the patient has made the initial diagnosis choice, the icity response is: Recommend 3 kinds of treatment plan, namely the treatment drug recommendation, the tumor hospital recommendation, the tumor expert recommendation. Icity also linked 3 different "ontologies", and the user made an association selection, such as choosing a cancer hospital. Then, the Chinese Academy of Medical Sciences Cancer Hospital, this time icity to do is logical reasoning, is the global to individual reasoning, because the Chinese Academy of Medical Sciences Cancer Hospital is a cancer hospital.

Medical Services: After the user chooses the cancer Hospital of Chinese Academy of Medical Sciences, Icity provides all the medical services of the hospital, including: Registration appointment procedure, medical cost assessment, intelligent navigation, and meteorological travel advice. The patient then makes a related operation, such as the "Registered appointment procedure", to complete the 3rd time human interaction.

Data Link: computer provides a complete set of procedures for online registration of cancer Hospital of Chinese Academy of Medical Sciences.

In general, a deep query is a system that begins by describing a phenomenon, symptom, or impression of something, and provides an alternative concept or entity that is related to a phenomenon or symptom through human-computer interaction. After the user makes the choice, the system based on the ontology and the related knowledge map stored in the cloud computing environment, in the gradual in-depth interaction with the user, through inductive conjecture, logical inference and related operation, finally provide the user needs, the online can be found, and improve the status quo associated with all the knowledge and data. From this point of view, the famous Turing Test is a special case of deep query. In summary, the underlying theoretical framework for deep query inference and manipulation should include:

1. Inductive guessing: attributes to Entities;
2. Correlation calculus: Inter-entity selection;
3. Logical reasoning: generality to instance;
4. Data Link: Entity to data (knowledge card).

The solution to the "deep query" service should include the establishment of a meta-language model describing various unstructured data, that is, the construction of an integrated representation and organization of data, the construction of unstructured data clouds containing ontologies, entities, knowledge maps and knowledge cards, the establishment of links between entities and raw data, and the realization of inductive conjecture, associated operations, Logical reasoning about the computational framework of knowledge and inference mechanism.

In the age of big data, the feelings of the human senses have been replaced by the big data produced by modern sensors. Therefore, the task of law discovery is to refine the mathematical equation between the predicate and the predicate by refining the basic concept and the precise mathematic relation between the concepts from the big data received by the sensor. In the context of social science and the big Data of human society, the relationship between these concepts can be causal or relational.

The decision generation from the intelligent traffic of Tian Tong Yuan

Let's use an example to illustrate what is the decision generation for smart cities.

Beijing Tian Tong Yuan area is about 8 square kilometers, the population is 915,000. The early peak period about 140,000 people to take the bus into the city to work, and the main road to the city only 2, one is through the Huilongguan and then bypass G6 Expressway to town, the other is the first through the soup road to Anlillo into town. The problems to be solved are: public traffic congestion, overloading of buses, poor safety and comfort, long time spent on roads.

Tin Tong Court Actual bus demand is: if each car 70 people, each ride time not more than 40 minutes to arrive at the destination, from 6:00 to 9:00 need an average of 10 buses per minute, a total of about 2000 trips, the length of nearly 40 kilometers.

In the current digital urban environment, the results of large data collection and statistical calculation are: Residents travel to Beijing 38 regions, but only 9 regions are the main destinations, accounting for 83% of the total number of passengers.

Ask the question: if the user to Baidu input: "Tian Tong yuan area work time, traffic congestion, by bus, into the city, to seek solutions", Baidu now provides services, but also can not give each work of the user specific, actionable advice, can not be said to be intelligent services. The intelligent answers that people have in mind should consist of the following steps:

decision-making qualitative: when the user enters "Tian Tong Yuan morning traffic congestion, seeks the solution", hoped obtains each other to give both "strategically advantageous position" and "conforms to the logic" the reply. So, Icity's answer is a question: "Contingency plan" or "long-term plan"? This is an inductive inference, as this is the behavior from the specific scenario to the scheme type.

When a user chooses "contingency plan", he or she makes a correlation operation or association inference, and Icity is associating two different types of solution. This association is provided by the computer, and the choice is made by the user (person), which is the first interaction with the computer.

proposal: After the user makes the choice of scheme classification, the icity response is: Propose 4 types of emergency solutions for users to choose, including "Bus scheduling optimization", "Hot Zone Direct", "Ride Time Optimization" and "Bypass route arrangement". Icity The 4 contingency plans, and the user makes one more choice, such as: "Ride time Optimization." Complete the second man-machine interaction.

Option : after the user chooses "ride time optimizes", because in the big data and the cloud computing environment, the Internet has the tin Tong Yuan 140,000 urban commuters each person's name, the social security number, the address, the mobile phone number, the current GPS address, the work unit address, All necessary information and data such as working hours statistic record, and also have the GPS address, driving route and the status of the passengers in the car in Tian Tong Yuan area. The data and the relationship between them constitute a dynamic planning system, and icity is to solve the problem of dynamic programming of big data. In the cloud computing environment, icity can directly or request the user to choose the solution method, call the Solver, and get the feasible decision. This includes: Every office worker will receive Icity's personal travel advice from the phone, such as when to take the bus and transfer advice. When the user agrees, Icity will also send the work adjustment time information to the work unit, in order to adjust the day, off-hours. Icity will also inform every bus driver about the bus route and time of the buses in equilibrium.

From the above example, we can draw the following about the icity decision generation of a macro description: Icity for the user's requirements and problems, through human-computer interaction in the ontology, entities, concepts, attributes, such as multi-level inductive conjecture, logical inference, related operations, in the interaction with the user process, revealing the nature of the problem, Provides all decision types that are relevant to the issue. After the user makes the choice, icity determines the problem solving mathematical model which satisfies the user's requirement, and determines the boundary condition and initial value to solve the model through the step-by-stage human-computer interaction. Then, icity call the solver, according to the user needs, provide the statistical law of the group and the dynamic evolution of the group, and according to the individual requirements of each group, give the individual solution, to achieve group balance, individual optimization. After the implementation of the group and individual solutions, icity will also be in real-time to perceive and verify the implementation of the solution, timely dynamic adjustment.

Unified Data Model is imperative

The city's big data is a digital image of the urban population's natural and social knowledge and information, collected through a variety of digital sensing devices, from all corners of the city and from different populations. They can be structured (such as traditional database files) or unstructured (such as voice, picture, video, etc.) and come from different historical stages. Due to the imbalance of social development, the format of raw data obtained from various ways is not uniform, which increases the difficulty of data processing and reduces the efficiency of data processing.

It is imperative to establish a unified data model in the historical stage of building icity. This data model should have the following properties: First, uniformity, that is, it can describe various structural and unstructured types of data in a uniform pattern, and the second is integration, that is, to directly inherit and adopt existing mature processing techniques on various types of data; Thirdly, it is related to the affinity, it supports multi-source Data feature association operation; That is to support data scale expansion and maintain processing efficiency; Five is evolutionary, that is, the evolution of data is recorded in a text sequence; Finally, the sensor-friendliness, that is, the data model can be embedded in a variety of sensors, the sensor from the outside world to receive raw data storage format. The advantage of this is that the potential of digital sensor can be fully exerted, and the processing efficiency of unstructured big data is greatly improved.

Icity is a group of software engineering

The city is a complex social ecosystem, its composition and evolution of the basic characteristics are: In the "macro-level", the municipal management institutions through the formulation of policies and the issuance of laws and regulations to the urban planning and allocation of resources to macro-control; on the "micro level", Through competition and market mechanism, citizens play a fundamental or decisive role in urban resource allocation and urban development. From this point of view, icity is a complex information ecosystem, and its big data at every moment is the image of the urban social ecosystem, which reflects the macro-control of municipal management organization to the city, as well as the digital description of the role of citizen competition and market mechanism to the allocation and development of urban resources.

If icity is used specifically for its software system, then group software engineering should be the preferred engineering method for the development, development and maintenance of icity. If we call Microsoft's method of developing Windows system a traditional software engineering method, the traditional software engineering method will face severe challenges in developing icity.

In Windows Vista, for example, the software development workload is about 60,000 modules, total code volume of 60 million lines, 9,000 professional developers, time spent 5 years to complete the development of the task. However, this traditional software engineering approach is severely challenged by Icity, because the overall scale and development effort of the latter will be far greater than that of Windows vista!

In recent years, Apple's App Store and Google's Android Market have brought us new insights that inspire social groups to participate extensively in software development. If you use the App Store or Android Market mode to invest 700,000 people, icity will be built in about 5 years! These two historic events in the software development industry tell us: Using Open source code, select crowdsourcing this group development model based on market competition will be a correct choice to solve the problem of the development, maintenance and evolution of the super-large software ecosystem such as icity.

To sum up, the main task of the intelligent transformation of Digital city is to provide "deep query", "Rule discovery" and "decision generation" 3 kinds of application services, therefore, it is necessary to promote the unified model of unstructured data from the digital sensing stage, and adopt the method of group soft engineering parts, Start with a more mature industry in digital cities as a pilot.

(This article is reproduced in the "Global Finance" magazine, the New Vision of ICT edition)

(For more information about Huawei, please follow Huawei's developer community, Huawei's own open door: http://developer.huawei.com/cn/ict/, don't ask me what I call, others call me Lei Feng )

From Digital city to Smart city