果冻传染煤九一

Dignitaries,distinguishedleaders,dear colleagues, ladies and gentlemen：

First of all, I would like to thank you organizers to invite me to this intelligent society forum as an occasion for world AI conference. I am very honored to be invited to give this talk on “The Intelligent City: A System of Systems”. I would like to divide this talk into four parts, first of all, I would like to give some basics of a system theory for iCities for intelligent cities, second section will be about digital twins based on Artificial Intelligence and intelligent agents, and then the third section I will give some examples of rational digital twins as models for an iCity system of systems, and finally I will give a brief conclusion.

If you look at the European Union, and how the European Union defines smart cities functional systems, they are very much the same as in China of course. And the extent to sustainable urban mobility, to sustainable districts and built environment, integrated infrastructures and processes in energy, information and communication technologies and transport, and the focus on citizens of course, the focus on policy and regulation, integrated planning and management, on knowledge sharing, on baselines, performance indicators and metrics, open data governance, standards, and business models, procurement and funding. And this not stated too much to say that this is really a very very nice collection of systems which together make up a vey good system. So, a smart city as far as intelligent city of course really is a system of systems. And if you look how to deal with it, and you have to deal with it in our daily lives, you have to deal with it in respect to governance, you have to deal with it in respect to transportation and many other issues. So it’s really a question, what’s the basics of an important intelligent information society where we have so many systems which are somehow interconnected.

And it’s the German Philosopher Niklas Luhmann who worked out the idea and this systems theory focuses on the three topics: Systems Theory as Societal Theory, Communication Theory, and Evolution theory, and the first two items will be the important ones for this talk. Luhmann states that a system is defined by a boundary between itself and its environment, dividing it from an infinitely complex or exterior system. The interior side of the system is thus a zone of reduced complexity where communication within a system operates by selecting only a limited amount of all information available outside. This process is also called “reduction of complexity” and elevates some available information in the environment to available knowledge within the system. The criterion according to which information is selected in the interior and processed is meaning. Both social systems and psychic systems, Luhmann’s, for instance, operate by processing meaning. And social systems are defined by Luhmann by recursive communication, so communication really builds up social systems. Modern society is defined as a world system consisting of the total sum of all communication happening at once. Individual function systems (such as the economy, politics, science, art, the media) are described as social subsystems which have “outdifferentiated” from the social system and achieved their own operational closure. Based on these insights, we can look at the Intelligent Information Society as a system of systems where humans and things are connected through subsystems and communicate with each other at the semantics, at the meaning level.

Just to explain that by some pictures so if we have the social system, we have a lot of information going at least to the outside of the social system, and if we have some filters, then we can have some reduction complexity by meaning, so these filters give meaning to the information. And that may mean that they really establish knowledge within the system. And not everything from the outside is being transferred into the system so we only have a few amount of knowledge within the system.

And now the question is how can twin cities people communicate with each other within the systems. And that of course can happen if it adheres to the same meaning of words they are using for their communication. So, the meaning is very very important for two persons and of course also for many persons communicating with each other within such a system. So, we have established the communication mechanics, at least halfly, within the subsystem. And now we would like to look at the system of systems and iCity is certainly a system of systems because we have interconnected subsystems like healthcare, smart meter, smart phone, smart home, smart building, smart grid, smart mobility, smart industry, smart devices and many many more. And for each of those subsystems, it is really that within the subsystem, there must be established meaning for the communication within the subsystem, and if some communication has to happen between subsystems, they also have to establish some common meaning through the communication content.

So the iCity is a system of systems for the information society, it is a complex communicating system of systems with many functions and cross-cutting interrelationships between them. And of course people somehow managed we have to digitize the iCity and to digitalize supports to support the city administration, the city management to improve the life of the citizens, and also the decision making for the planning of future iCity developments. So, this iCity digitalization enables models, each model of the current and future dynamic behaviors of an iCity, and the Digital iCity Twins offer holistic solutions for modeling the life cycle of iCity processes and services beyond the current partial and disconnected iCity models.

So in the next section I will very briefly introduce how AI can help to modern those systems of systems. And if you look at the simplest case as CyberPhysical AI systems which reflect the digitalization of the real world as its digital twins, we can very easily see that on the one hand side, there is environment, and on the other hand there is a system and this system has sensors to perceive information from the environment and also to act on the environment of course for its actuators. So this system is connected to the environment and it somehow models environment. It’s a digital twin and it’s an intelligent digital twin because it usually can auto-reason about environment, reason about its goals, can process information and can even make decisions. So we really have known the connection between the physical world and the virtual world, and that means we also have to look how the decisions in the virtual world will be taken and fortunately there is rationality of AI systems, and the rationality of AI systems is defined as the ability to choose the best action to take in order to achieve a certain goal, given certain criteria to be optimized and also of course the available resources.

So we have known the connection between the physical the real world and the virtual world, and the question is now how can be modeled that CyberPhysical system so it becomes a digital twin of some part of the real world, and also of course apply to iCity planning and management dynamics. And fortunately, there is something in artificial intelligence which is called multi agent systems and it can also have Multi Agent Based Simulation which differs from other kinds of computer-based simulation in that (some of) the simulated entities are modeled and implemented in terms of agents. So we have CyberPhysical systems as agents and we can digitalized iCity systems by CyberPhysical Systems so they become Digital iCity Twins. Digital iCity Twins are implemented by AI agents with rational, cooperating, and proactive decision making. And Goal-oriented Digital iCity Twins model the dynamics of an iCity by simulation models based on their local semantics knowledge and on negotiations between them. So we have gain here the meaning the local semantics knowledge within the system and also the communication which is the negotiations between agents. And the global system behaviour then emerges from the combination of local system interactions by Digital iCity Twins and it’s also possible to predict the consequences of management and/or planning decisions by those digital twins and by the simulation incorporating digital twins.

So we have multi agent based simulations which digitalized city twins for iCity planning and management decision support. And if you look at the data beneath forth that, that could be for instance, historic satellite image data, it could be homologous case data of the past of the city, of course local land use maps, points of interest and planning assumptions. And what are we going to do with these data? We apply machine learning algorithms so we generate knowledge from the past and we detect iCity system patterns which enable us then to define agents, for instance, government agents, investor agents, planner agents or citizen agents. And then we can look at its scenarios for instance, its scale of development, its speed of growth, its trends of development direction, its function changes of the city, and economical growth of the city, and of course also its environment protection cases in the city and many many more. And when we establish those agents, then we can simulate the iCity scenarios on top of simulation platform, and based on simulations like that you can predict future decision consequences and of course depending on the input we collected earlier. So we can really predict future iCity scenarios and can see what consequences the earlier decisions could have. So we have made the connection now between the real city and the intelligent city and how to model also the dynamics of intelligent cities.

And now I just would like to give some examples on rational digital twins as models for an iCity system of systems. This is an example for urban population prediction where we at Tongji University and the CIUC look at four age groups: juvenile, youth, midlife, and elderly. And this is a simulation which goes over many years and of course it simulates dynamic process and maybe you can see that a lot of people will be in dynamic life part of their lives. This is another example for future district development where we simulated with agents for the government, for citizens, for planners and also for investors. And for instance in this context we started at 2018 and then we simulated the course of that district and until 2030. And you can see this district very quickly builds up and that there will be a proper distribution of leisure areas, industry areas and so on. Next example is the prediction of energy consumption in the district. And you can see that we on the one hand can simulate how much energy will be used within a year and also of all the years from 2018 to 2035. What the water consumption would be within the time phase, what the amount of waste generation would be and also what within that phase will emit carbon. So this is also another simulation for the future based of course on currentdata and on planning assumptions. This is an example I would like to give today and it’s the installationof public services facilities, the prediction of them for the industries. And it’s also covering the time phase from 2018 to 2035, and you can see that it’s very easy to see that a lot of additional facilities have been introduced in the communities. So far examples.

And now I would like to give some conclusion. And I hope that I could show that we already have mature technologies for iCity System of Systems decision support. We have big data streams analytics for the digitalization from data to knowledge, we have CyberPhysical systems as the technological basis for digital iCity twins, we have iCity semantics as a basis for an end-to-end integration of the iCity systems of systems models for iCity processes and also for the establishment of international standards, we have AI methods for digital modelling and simulation including the integration into existing domain-specific modelling methods and tools, and we have also holistic models for the iCity highly interconnected systems of systems including dynamic intelligent services, and finally it’s also possible to validate iCity processes for iCity planning phases and for the iCity management.

This is what I would like to talk to you about today, 谢谢大家, and I thank you very much for your attention.

(Thanks for the translation provided by HE Yifan, a postgraduate of Tongji Law School, SU Miaohan, an Associate Professor, and WANG Fengge, a teacher of the School of Architecture and Urban Planning.)