(1) Uncertainty, complex systems and the nature of planning

The essence of planning is the guidance and regulation of the future. Under the conditions of the market environment, the quality and quantity of land use included in the future will always change with time and space. Especially with the rapid development of regional economy and the rapid advancement of urbanization, it becomes even more complicated. . As a result, people are bound to be full of uncertainty when taking actions in their environment or system. From a planning perspective, there are at least four types of uncertainty: environmental uncertainty, value uncertainty, uncertainty in related decisions, and uncertainty in solution search (Hopkins, 1981).

Traditional planning's handling of uncertainty is mostly based on an ideal problem framework, that is, similar to the small world proposed by Savage (1972), where utility is constructed through utility and subjective probability. According to the theoretical theorem, the decision-maker can choose the best action calmly and rationally to maximize the decision-maker's utility. However, planning practice and cognitive psychology experiments have shown that people's decision-making behavior usually violates the criterion of utility maximization, and traps will appear in the decision-making process, such as anchoring, status quo, and sunk costs ( sunk-cost) and evidence (confor-ming-evidence), etc. Land use covers the three major industries and 16 industries of the national economy, making it a very complex system. Due to limitations in information processing capabilities, people often simply perceive land use as a tree-like hierarchical structure, but in fact the system is a semi-lattice structure (Alexander, 1965); another example is that most people think It is believed that the organizational structure of planning management is hierarchical, but in fact the organizational system of planning management is extremely complex and its evolution is full of uncertainty. Therefore, traditional planning methods often lose their effectiveness in solving practical problems due to distorted images of complex systems based on limitations in cognitive abilities.

Taking a city on the eastern coast as an example, between 1997 and 2005 when the last round of land use master planning (1997-2010) was implemented, the actual area of ??new urban construction land within the planned urban construction land boundaries was 29.1 km2, and the area of ??new urban construction land outside the boundary is as high as 33.9km2. In addition, judging from the spatial distribution of construction land outside the boundary, a large amount of new construction land is located at the outer edge of the planned urban construction land boundary. A total of 66.3km of planned urban construction land has new land development at the outer edge of the planned urban construction land boundary. , accounting for 28% of the total border length (239.3km). In 2003, the city began a new round of revision of the land use master plan (2006-2020). Due to various reasons, the plan lasted for more than seven years and is expected to be approved by the State Council in the first half of 2010. What is worth thinking about is that during these more than seven years, the plan has been constantly revised. In total, nearly 50 versions have been revised, with an average of one revision every two months. This fully demonstrates that in the current rapidly developing complex system environment, planning is a multi-stage dynamic decision-making problem. These decision-making behaviors with interdependence, irreversibility, indivisibility and imperfect foresight form a complex system through interaction (Hopkins, 2009) .

Faced with planning problems that require dynamic decision-making, complexity theories and methods provide a basic tool for recognizing the overall phenomenon produced by the individual interaction of each element in the system and working to solve practical problems. From chaos, fractal, nonlinear dynamic systems, artificial life to complexity theory, the knowledge system of complex systems is constantly improving. According to the core definition of complex systems, the high complexity of land use system evolution, or the ill-defined problems that planning must solve due to environmental uncertainty, all come from the interaction between different decisions in the system. The intricate relationships can be understood under the theoretical framework of complex systems. Relying on the superiority of computer processing power and looking for cognitive causes of uncertainty from existing literature on the limitations of information processing capabilities (such as the capacity of permanent memory and temporary memory and the time it takes to convert information between them) , the knowledge block of complexity theory can suggest appropriate ways to deal with planning uncertainty problems.

(2) Planning and decision-making paradigms under complex systems: framework rationality, opportunity flow and spatial trash can theory

Land use is a complex system, and the core of planning lies in exploring how to Make rational solution choices in this complex system. The completely rational choice theory in economics is not enough to deal with complex systems such as land use.

Currently, the most widely accepted paradigm of rationality is subjective expected utility theory, which assumes that the world facing decision makers is simple and emphasizes the importance of making single, independent decisions. As everyone knows, when the world that decision makers face is complex, this kind of thinking will fail. In addition, due to the limitation of the decision-maker's own ability, he cannot be completely rational in the decision-making process, but makes decisions under the premise of "bounded rationality". Therefore, the subjective expected utility theory has been severely challenged by psychologists and experimental economists in recent years (Hogarth et al., 1987). Choice theory that divides decision-making paradigms into descriptive, normative and prescriptive is not helpful in solving practical problems (Lai Shigang et al., 2010). Through exploration and combing, we proposed framed rationality to solve the uncertainty problem of land use planning in complex systems, which may be a new cognitive path.

In frame rationality theory, frames are defined as decision-making events under the behavior of the decision-maker. According to the psychological experiment designed by Kahneman and Tversky in 1979, the framing of the problem will affect the choice situation perceived by the decision-maker and produce a preference reversal phenomenon (Kahneman et al., 1979). Using the same questions from Kahneman and Tversky's experiment, we found that a statistically significant number of subjects made choices that maximized their subjective expected utility regardless of how the question was framed. In other words, preference reversal does not violate the subjective expected utility (SEU) model, but rather verifies the validity of the model within a specific framework. Therefore, it is believed that no matter how the problem frame is defined, decision makers are generally rational as defined by subjective expected utility theory, and this explanation of choice behavior can be called frame rationality. Frame rationality denies the assumptions of neoclassical economic theory and the concept of comprehensive complete rationality developed from positivism (positivism) scientific philosophy, thereby consolidating the validity of the SEU model (or similar concepts) in a specific framework. Frame rationality theory no longer searches for the optimal action plan in planning, but explores which subset of planned actions can best demonstrate the robustness of benefits under all possible future scenarios. This concept is similar to the coherentist theories of planning proposed by Donaghy et al. (2006) in response to the impossibility of a complete general planning theory.

In the process of preparation and implementation of land use planning, the current method is still guided by the theory of subjective expected utility. Its typical feature is to assume an ideal future and look for optimal actions; and believes that The optimal allocation of land use under specific goals can be achieved through strict top-down control. In complex systems that are closer to the real world, due to the dynamic flow of decision-making elements and rapid changes in the environment, there are also large differences in information, options, and goal orientations between the central and local governments. In land use planning formulation and subjective expected utility theory, The assumed completely rational decision-making environments vary widely. Therefore, the optimal resource allocation guided by subjective expected utility theory cannot be realized.

Framework rationality provides a thinking paradigm for rational selection of planning plans, but it does not explain how planners find paths and plans for efficient allocation of land resources in the complex environment they face. Therefore, constructing a paradigm is an important issue worthy of serious discussion. Because effective paradigms can make problems transparent, and then discover effective solutions. The stream of op-portunities model proposed by Professor Hopkins aptly describes the real decision-making situations faced by planners. Based on the concept of the garbage can model (Cohen et al., 1972), he explained that when planners face complex and uncertain environments, they should grasp the decision-making situation in the stream of opportunities and use appropriate plans to make decisions. Problem solving (Hopkins, 2009). Based on the trash can model, Lai Shigang proposed the spatial trash can model by taking into account the spatial factors of the location. He believes that elements such as specific decision makers, solutions, choice opportunities, problems, and facility locations randomly meet in the opportunity flow to generate decisions and then solve the problem. The experimental results he designed showed that the main effect of channel structure is statistically significant in affecting system performance, but the impact of spatial structure is not significant (Lai Shigang, 2002). This means that in the process of land use system evolution, the traditional method of improving system performance through spatial design is not as effective as changing activities through institutional design, or at least a combination of both. In many cases, land use systems may be disordered and the causal relationships between elements in the system are not intuitive. Planning solutions are sometimes generated before land use problems arise, and planners can only continue to plan and solve problems in such an environment to achieve the goals of land use planning.

The opportunity flow paradigm or the spatial trash can model both indicate that the complexity and dynamic changes of the land use system are not under the control of planners. The only thing planners can do is to understand the relationship between decisions, problems and plans in time and space, and continuously Formulate plans, revise plans and implement plans.

(3) Planning model selection in complex systems

Planning that faces uncertainty and considers relevant decisions traditionally includes design (design) and strategy (strategy) Two planning mechanisms (Hopkins, 2009). Design is the process of finding an optimal solution to a linear programming problem. When faced with the problem of land use planning for which it is difficult to find the best solution, design falls back to the second best and can only become a means of seeking local optimization. Strategy is different from design. What is pursued is not a one-time solution, but expedient measures. It is closely related to decision analysis and is most suitable for environments where many decision makers are involved and face great uncertainty, such as complex systems. Both design and strategic approaches to planning will bring net benefits to planners, but the timing of their application is different. Design is like a comprehensive long-term plan. Once formulated, it needs to be implemented according to the plan. Strategy is like a short-term rolling plan that is constantly revised and formulated over time.

Previous computer simulation experiments have found that although design planning with the goal of optimization is more effective than no planning in resource utilization and brings order to the system, it cannot solve more problems ( Lai Shigang et al., 2009). When faced with regular or stochastic systems, such as economic systems that have reached equilibrium, design may be able to meet the needs, because the occurrence of events in these systems presents a fixed pattern; but when faced with complex systems in between, When, for example, land use, events occur in unpredictable ways, designs that consider decisions independently will fail. At this time, strategic planning that considers relevant decisions can lead to better benefits. Strategic planning is particularly more effective than design planning in solving problems arising from rapidly changing environments. Mainly because strategic planning is flexible and its formulation costs are low. From a certain perspective, strategic planning is between progressive no-planning and blueprint-based design planning, which is what complexity theory says is between chaos and order, so it can take into account the advantages of both (Han Haoying et al., 2009) . A comparison of the main features of design and strategy is shown in Table 7-1.

Table 7-1 Comparison of characteristics of design and strategy

(According to Hopkins, 2009)

(4) Implementation of strategic planning model Path

Traditional Chinese planning, whether urban planning or land use planning, is a blueprint-style plan that determines the complete results of relevant decisions at one time, so it is a typical design-based plan. This planning model mainly consists of a set of highly interrelated actions, considering only situations where a few actors are involved and the uncertainty of the actions is small. Therefore, it is suitable for planned economies with a single development entity and relatively certain development behaviors; under market economy conditions with multiple entities and sporadic development, its role is limited. Today, as land use becomes increasingly complex, it should be a general trend to transform from the traditional design-based planning model to a more flexible strategic planning model. This transformation can be achieved by the following four changes: ① From "time-driven" to "event-driven"; ② From "result-based" planning to "process-based" control; ③ From single-mode zoning to "event-driven" Diversity zoning transformation; ④ Transformation from “zoning-type” planning to “zoning-permit hybrid” planning (see Table 7-2).

1. Transformation from "time-driven" to "event-driven"

Planning control usually includes two basic methods: time-driven (time-drive-en) and event-driven Event-driven. Time-driven planning means that the decision-making time point for land use expansion is fixed, such as being revised every five years; while event-driven planning means that land expansion decisions occur when the land inventory decreases to a certain threshold. China's previous master planning preparations, whether urban master plans or land use master plans, were all traditional time-driven plans. For example, the land use master plan is theoretically revised every 5 years, or 10 years, or 15 years.

Table 7-2 Comparison between design-based and strategic planning

Knaap et al. (2001) analyzed the holding cost (holding cost) of land stock management in urban land expansion, Theoretical estimates were made on order cost and deficiency cost, and it was found that although the traditional time-driven planning revision method saves administrative costs, it also requires land costs because it cannot be flexibly adjusted according to the actual development status of the city. Environmental costs such as rising housing prices and deteriorating environmental quality caused by insufficient inventory are likely to reduce the overall benefits of spatial planning. The advantage of event-based planning over time-based planning is that its decision to expand urban land stock will not incur additional loss costs beyond the retention and ordering of land stock.

The shortcomings of the traditional time-oriented planning control method have been confirmed by many recent domestic research cases.

For example, in the field of urban planning, Mao Jiangxing et al. (2008) discussed the regulatory effectiveness of Shenzhen urban planning on land use based on the data released by the construction land planning license, and found that the scale and spatial layout of urban construction land are related to the urban master plan. A greater degree of deviation. Tian Li et al. (2008) compared Guangzhou's urban master plan (2001-2010) data with 2007 urban spatial development status data and found that the implementation effect of the urban master plan was not ideal. In their study of Beijing, Han Haoying et al. (2009) used multi-temporal remote sensing to test the effectiveness of planned urban construction boundaries (UCB) in controlling urban land growth, and found that the actual differences between the 1983 version of the plan and the 1993 version of the plan were During the implementation period, the actual growth scale of urban construction land outside the UCB within the Sixth Ring Road of Beijing is higher than the growth scale within the UCB; moreover, the developable land stock within the UCB initially set in the master plan cannot satisfy the actual land development needs. In the field of land planning, time-driven control of land use is also far from meeting the requirements of planning objectives. The existence of these problems shows that it is necessary and urgent to establish event-oriented strategic planning in China.

2. Transformation from “result-based” planning to “process-based” control

The transformation from result-based planning to process-based control has been widely discussed in the field of urban planning. The former is emphasized as the core content in traditional design-based planning, that is, by drawing a development blueprint for several years, using control indicators and spatial forms in different periods such as the short-term, mid-term and long-term to control urban development and achieve planning goals; The latter does not set control indicators at a certain future point in time or the ultimate form of the plan, but details and clearly defines various requirements in the planning implementation process, including the conditions for development permission, development implementation and implementation. The conditions that need to be met, as well as the form and degree of public participation, therefore fall more into the category of strategic planning.

After the 1960s, the United States gradually developed the advocacy planning model and the liaison planning model based on the traditional rational planning model. Land use planning has gradually shifted from an exclusive professional field dominated by the government and experts to an arena where multiple stakeholders participate and compete (Kaiser et al., 1995).

In order to achieve good planning, it is important to control both "results" and "processes". However, the current practice of land use planning places too much emphasis on results and ignores the process. Therefore, this book believes that the focus of land use planning control should change from "result-based" planning to "process-based" control. In order to achieve this transformation, it is necessary to reform the planning indicator system and preparation methods. Specifically, the plan may not clearly stipulate the scale, development sequence and spatial form of each land use zone at several future time points, but set access indicators for each project, such as the amount of land consumed for each new population. , whether it complies with national industrial policies, unit land investment, floor area ratio, building density, number of jobs it can provide, tax revenue and environmental impact, etc. For process control indicators, on the premise of obtaining full understanding and supervision of the local public, dynamic adjustments can be made according to local social and economic development conditions and special needs to cope with uncertainties. The core of process-based planning is that the process is reasonable, not the results. Therefore, the main investment in planning can be shifted towards the design and control of processes rather than the setting of outcomes.

3. Transformation from single-mode zoning to diverse zoning

After the United States established its land use planning and control system, it has been adjusting and improving it. Take zoning (zon-ing), the main land use control method in the United States, as an example. Traditional Euclidean zoning is too rigid and difficult to achieve some important planning goals, such as leaving enough space for uncertain development. , provide affordable housing for low- and moderate-income people, and reflect the interests of local residents and groups, etc. Therefore, on this basis, various types of new zoning models have emerged, such as inclusive zoning, incentive zoning, performance zoning and negotiated zoning ( Proceedings of American and Japanese Urban Planning Experts, 1993).

(1) Inclusive zoning is a zoning regulation that requires or encourages housing developers to provide a certain proportion of housing to low- and middle-income families. This type of planning attempts to incorporate social equity goals into land use planning. Inclusive zoning can be mandatory or voluntary, allowing developers to build higher density than zoning regulations in order to achieve higher profits, but at the same time must set aside a certain proportion of the excess housing as low-cost housing. In some cities, developers can choose to pay into affordable housing construction funds instead of building such housing directly.

(2) Incentive zoning requires developers who provide low- and middle-income housing to be rewarded with building area. For example, in 1987, New York City passed a zoning resolution allowing incentive zoning in high-density areas.

Developers who provide new or preserve existing low-income housing on beneficiary building land, within the same community, or within half a mile of a beneficiary building can receive an incentive of 20% of the additional floor space.

(3) Performance zoning is a flexible zoning form with a "score system" as its core. It only focuses on the impact of investors' use of land on neighboring areas. It does not care about the land use properties and other attributes of the proposed project. Performance-based zoning stipulates that a proposed project will receive a certain number of points for a certain advantage and deduct points for a certain disadvantage. Once the project has been designed and the total score exceeds the prescribed standards, it will be approved for construction. Regardless of the type of land use for the investment project, even a factory can be approved if it meets the zoning standards. The purpose of performance zoning is to eliminate any negative side effects that development projects may have on adjacent areas and let the market determine the most appropriate use of the land.

(4) Negotiated zoning stipulates that local governments can negotiate with individual investors to resolve real estate zoning issues in detail one by one. The most common method is "planned unit development (PUD, or planned unit development)". The theory is that the government should be able to negotiate some compromises on its zoning objectives in certain areas in exchange for investors benefiting from the public welfare. Make a contribution. This form of zoning does not apply to the entire area, but is determined by the government in consultation with the adapted investors for each land zoning.

China’s land use planning can draw relevant experience from zoning changes in the United States. On the one hand, planning can combine land policy with other policy requirements such as housing and transportation to achieve more comprehensive and diverse goals. On the other hand, planning can set goals that can be achieved through market-based operations while ensuring fundamental requirements, such as reducing the side effects of development on neighboring areas and achieving more social benefits, using a "score system" and "negotiation." "System" and other flexible methods to enhance the operability and implementation effectiveness of the plan.

4. Transition from "zoning-based" planning to "zoning-permit hybrid" planning

Zoning-based planning and permission-based planning are two different types of land use control models. The former is for the government to make comprehensive and detailed regulations on the nature, development intensity and other indicators of land use within a certain range in advance. If a certain land use meets all the requirements specified in advance in the plan, the planning department will approve the development and utilization. The latter does not clearly stipulate all utilization requirements in advance, but applies for approval on a case-by-case basis, gradually clarifies planning requirements during consultations, and approves based on the finalized requirements.

From a global perspective, the United States is the main country that implements zoning planning, while the United Kingdom is the main country that implements permit planning. Considering the characteristics of implementation, development permit planning is more flexible than zoning planning, but its information collection costs, property rights division costs and other transaction costs are higher than zoning planning (Lai Shigang, 2002), so it is more time-consuming and laborious.

China is currently in a stage of rapid urbanization. Without significantly increasing planning technology, controlling a large number of incidental development activities is the primary task of current planning. Zoning-type planning, a "check-permit" planning technology that covers a wide scope, is simple, easy to implement, and low-cost, and it can be regarded as a more effective method (Tan Zongbo, 2001). However, this planning method can only enable land development to meet some basic requirements, such as that polluting industries will not affect residential life; it cannot do anything for higher needs, such as creating a good spatial form and living environment. Permit-based planning can make up for this disadvantage and achieve better planning results through case-by-case negotiation; however, because it is time-consuming and labor-intensive, and relies on good consultation and the establishment of a public spirit, it cannot be promoted on a large scale in many places. In practice, the two can be combined, such as adopting a zoning-based planning approach in most urbanized areas that require general control, while adopting a permission-based planning approach in specific areas that require key development to achieve "point" and The combination of "face". Of course, in the context of China's current national conditions, permission-based planning may produce more rent-seeking, but this is a problem at the governance structure level, not at the institutional level.