The Future of Management: DAO to Smart Organizations and Intelligent Operations

In the future, management in smart societies will revolve around knowledge workers and the works they produce. This article is committed to explore new management framework, model, paradigm, and solution for organizing, managing, and measuring knowledge works. First, the parallel management framework is presented that would allow for the virtual-real interactions of humans in social space, robots in physical space, and digital humans in cyberspace to realize descriptive, predictive, and prescriptive intelligence for management. Then, the management foundation models are proposed by fusing scenarios engineering with artificial intelligence foundation models and cyber–physical-social systems. Moreover, the new management paradigm driven by decentralized autonomous organizations and operations is formulated for the advancement of smart organizations and intelligent operations. On these basis, the management operating systems that highlight features of simple intelligence, provable security, flexible scalability, and ecological harmony are finally put forward as new management solution.

Modern management theories in the west began with the industrial revolution, and its founder was Henri Fayol, who is a French mining engineer [10], [11], [12], [13]. The classic "14 principles of management" was put forward by him in "General Principles of Management" in 1908 to commemorate the 50th anniversary of the French Mining Society. In 1916, he published an article entitled "Industrial Management and General Management" in "Bulletin of the Mining Society," proposing his managerialism represented by "management elements" and five functions of planning, organizing, commanding, coordinating, and controlling (POC3), which paved the way to modern management science.
The goal of modern management is to explore a scientific and standardized way for improving the productivity. To this end, Taylor [14] put forward his principles of scientific management, which utilized Taylor's stopwatch to measure manual works, thus, solve the problem of monitoring the workload and efficiency of production tasks. The reason comes down to that he did not consider or even completely miscalculated the human nature and their response to the changes in production behaviors and productivity [15], [16]. In 1969, Peter F. Drucker introduced the concept of "knowledge works" in his monograph entitled "Age of Discontinuity," and he believed that knowledge works and knowledge workers were the breakthrough to change management and promote its transformation from industrial enterprises to intelligent organizations. Consequently, it is very crucial to shift from measuring manual works by scientific management to measuring knowledge works by smart management and also re-examine and deal with human nature and social factors of knowledge workers and their organizations on this basis [17]. Maslow's work of "hierarchy of needs" has been regarded as an important theory to explain human nature and motivations [18], but it does not clarify how to deal with humanity. The subsequent theories, such as "X theory," "Y theory," "Z theory," and "3-D theory" [19], [20], [21], have tried to deal with this problem while did not really appreciate the complexity and capacities of humanity. More importantly, they did not solve or even involve the primary problem of measuring knowledge works.
In smart societies, management systems are essentially cyber-physical-social systems (CPSSs) centered on knowledge workers and their works. It is very challenging to measure knowledge works and improve their productivity under the premise of fully considering and respecting humanity and rationality [22], [23]. This not only involves effectively dealing This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ with the intellectual behavior response of knowledge workers that may be induced by the productivity improvement but also seeking an organic and sustainable balance between personalized work satisfaction and productivity. Besides, the coupling of engineering complexity and social complexity aggravates this challenge [24], which cannot be coped by the traditional management methods. Therefore, it is urgent to explore new management theories and technologies.
Parallel management takes the complexity science that integrates technological science and social science as the foundational tool, and its basic idea, framework, process and technology systems are formulated for the management and control of CPSS [25], [26], [27]. Because it can effectively handle the coupling engineering complexity and social complexity in CPSS, parallel management has great potential to provide tools and methods for studying the humanity of knowledge workers as well as measuring and managing their knowledge works accordingly. The main contributions of this article can be summarized as: new management framework defined by parallel management is used to manage knowledge workers and their works, new management model, namely, management foundation model is proposed to measure knowledge works, new management paradigm based on decentralized autonomous organizations and operations (DAOs) is formulated to fully consider the humanity of knowledge workers, and new management solution referred to as management operating system (MOS) is designed to realize intelligent management of knowledge works.
The remainder of this article is arranged as follows. Section II presents parallel management as a new management framework. Section III proposes management foundation models as a new management model. Section IV discusses new management paradigm of decentralized management. Section V puts forward the MOSs as new management solution. Section VI concludes this article.

II. NEW MANAGEMENT FRAMEWORK: PARALLEL MANAGEMENT IN THREE WORLDS
Based on ACP (Artificial Society + Computing Experiment + Parallel Execution) theory, parallel management can realize the description, prediction, and prescription of the real-world via virtual-real interaction, closed-loop feedback and parallel execution of the physical world, the psychological world, and the artificial world [28]. Parallel management is a new management thinking that relies on social computing, knowledge automation, and artificial intelligence (AI) to provide tools, methods, and platforms for the production and measurement of knowledge works, the improvement of knowledge productivity, and the sustainable management of knowledge organizations. As follows, the unique features of parallel management will be introduced, and its basic framework with detailed processes and functionalities modes will also be studied.

A. Unique Features of Parallel Management
As a new framework to deal with the complexity of measuring and managing knowledge works, parallel management possesses the unique features of three worlds, circular causality, and parallel spaces. On the basis of new management worlds represented by three worlds, new management spaces represented by parallel spaces are built to realize the new management logic of circular causality driven by virtual-real interaction.
1) Three Worlds: The management world mainly consists of the physical world, the psychological world, and the artificial world (as shown in Fig. 1) [23]. The development of the physical world relies on the industrial technologies ("old" IT) to eliminate resource asymmetry, the development of the psychological world depends on the information technologies ("past" IT) to eliminate information asymmetry, and the development of the artificial world requires the intelligent technologies ("new" IT) to eliminate intellectual asymmetry. They must be used simultaneously in parallel to develop and manage our world. The artificial world is by nature the knowledge world, where the management innovation concerning knowledge works, knowledge economies, and knowledge societies will happen. We believe knowledge societies and knowledge organizations will be comprised of three types of participants, namely, humans with instinctive rationality, robots with adaptive rationality, and digital humans with computational rationality. Humans are responsible for leading the organizations; robots are intelligent machines that take charge of physical works; and digital humans are intelligent programs and information machines shaped according to humans and behave in human manner.
2) Circular Causality: Based on the generalized circular causality theory, three types of participants in three worlds are managed in parallel, forming a series of parallel causal events in each world. That is, different management decisions produce different decision results, and then the initial management decisions are modified according to the ideal decision results to produce a new round of causal events, thus, forming a specific opposite and unified management closed-loop feedback mechanism in each world. Furthermore, by integrating the unhappened "What-IF" thinking in the artificial world and the ongoing "IF-Then" experience in the real world, a new circular causal management logic of virtual-real interaction and closed-loop feedback will be established [23].
3) Parallel Spaces: Parallel spaces are built to realize virtual-real interactions and integration of the physical space (physicspace) and the cyberspace [25], [29]. Through the evolution and emergence of management systems in cyberspace, the description and prediction of management systems in physicspace can be realized, so as to guide management decisions in physicspace to approach and converge to management goals. This forms a management cycle from the unordered emergence in physicspace to the orderly convergence in cyberspace, and then from the emergence in cyberspace to the convergence in physicspace. In parallel spaces, humans are put in the leading position of parallel management, and humans, robots, and digital humans are deeply integrated and effectively collaborated to realize management intelligence.

B. Framework of Parallel Management
Parallel management mainly refers to the virtual-real interaction and parallel execution of physically embodied organizations and computationally embodied organizations. The basic framework, detailed processes, and functionalities modes of parallel management are displayed by Fig. 2. Physically embodied organizations represent the actually managed organizations, including the core enterprises and organizations in the management chain and ecosystem. Computationally embodied organizations represent one or more artificial enterprises organizations corresponding to physically embodied organizations, in the forms of digital twins, software-defined organizations, metaverses, virtual societies, etc.
Through the parallel interactions between physically and computationally embodied organizations, especially, the parallel drive of data (e.g., experiences) and knowledge (e.g., rules), and the parallel execution of virtual and real implementation, the double feedback and double closed-loop between virtual and real management are realized. Consequently, the generalized circular causality in three worlds that the future affects the past, the past affects the present, and the present affects the future is formulated. Parallel management has formed the complete decision-making process and operation platform system, which has three main functionalities modes.
1) Learning and Training: Parallel management adopts the decentralized autonomous online and offline management learning and training. The computationally embodied organizations are used as the center for humans and robots to carry out management learning and training synchronously.
Under the connected and combined artificial and actual cases or scenarios, management animation, interactive film, and television are used to rapidly improve the management decisionmaking abilities and enhance the leadership of managers.
2) Experiment and Evaluation: Parallel management takes computers and cyberspace as the social laboratory of management, uses management small data to generate management big data, and then refines management big data into management deep intelligence, that is, the personalized accurate knowledge or "small intelligence" for specific problems and scenarios. In addition, acceleration and pressure tests can be conducted for different management solutions, to calculate various statistic indexes that serve as the basis for selecting and implementing different management decisions.
3) Management and Control: Through the virtual-real parallel interaction, the precise intelligent management with the characteristics of virtual-real double feedback and double closed-loops is realized, that is, parallel management based on generalized circular causality and circular logicality. The future vision generates "history," the historical data generates the current scenario, and the current management guides the future performance, so as to make the operation of physically embodied organizations follow the desired states depicted by computationally embodied organizations as much as possible for reaching the management goals.
Parallel management will promote the future management to become a programmable ecosystem that can be described, computed, experimented, estimated, verified, compared, and executed. It not only focuses on each management chain but also supports the entire management ecology, and finally realizes sustainable management with descriptive, predictive, and prescriptive intelligence. Consequently, parallel management framework can be applied to the management in individual, intergroup, and social dimensions, and well serve knowledgecentered management, such as person-post matching, group collaboration and competition, social governance, etc.
As follows, the modeling, organizational, and operational technologies involved in the realization of parallel management framework will be elucidated in detail, which mainly covers scenarios engineering, foundation models, CPSS, blockchains, DAOs, smart contracts, etc.

III. NEW MANAGEMENT MODEL: FOUNDATION MODELS
FOR MEASURING KNOWLEDGE WORKS Knowledge works do not only reflect the intellectual capacity but also take the humanity and rationality into consideration. We believe that knowledge works of humans, robots, and digital humans will be the main productivity in the upcoming smart societies, and the production relationships will be built on the division of knowledge works. Consequently, the measurement or even quantification of knowledge works becomes particularly important in management. However, it is very tough and challenging to measure knowledge works, due to the lack of proprietary tools for managing knowledge production. The existing infrastructures in the modern industrial society, such as computers, software, tool kits, testing systems, etc., are too primitive to measure knowledge works.
In view of this, new management models integrating the generalized environment, algorithms and technologies that tailored to managing knowledge production under the parallel management framework must be proposed to design the "stopwatch" for measuring knowledge works with the purpose of improving management efficiency and effectiveness. The emerging intelligent technologies, such as scenarios engineering, foundation models, and CPSS (or metaverse), have laid the basic foundations. Specifically, scenarios engineering builds the trustworthy and interpretable management scenario, foundation models form a scalable and universal cognition, and CPSS extracts knowledge from data via these models and then form scenario-oriented management intelligence.

A. Scenarios: From Features Engineering to Scenarios Engineering
For the future of management, one of the most crucial objectives is to transfer the traditionally effective case teaching and training to intelligent and efficient learning and automatic testing, which calls for strong supports from the structured and ordered management environment and the trustworthy management models. Through feature extraction and the resulting features engineering, AI technology has achieved great development on generating useful knowledge. Features engineering contributes to transform data into featured knowledge that can be better applied into management, thereby realizing the data value and improving performance of management models [30], [31], [32]. However, the use of features engineering in management may bring in high-data collection costs and even some systematic risks due to its inherent defects, such as bad features will elevate the difficulty of management model construction as well as weaken its performance; the "black box" learning model results in bad interpretability and controllability; the lack of enough consideration on reliability, security, and sustainability makes management models and the resulting management decisions not trustable [33], [34], [35]. In an effort to eliminate these defects and make management models more trustworthy, scenarios engineering must be introduced into parallel management.
Scenarios engineering is a novel intelligent technology substantially distinct from features engineering that can enhance the management performance as well as improve the trustworthiness in management [36]. By virtue of the infrastructures of intelligence and index (I&I), i.e., cognitive intelligence, parallel intelligence, crypto intelligence, federated intelligence, social intelligence, ecological intelligence, as well as their corresponding evaluation indexes, scenarios engineering can achieve the comprehensive goals of future management. Based on these multidimensional indexes, calibration and certification (C&C) are used to identify and predict suitable models and strategies to get desired output [37], [38] and also grant the reliable certificates for the qualified contributors, thus, can ensure the quality and performance of management results. Aiming to make management more efficient and effective, the entire decision-making process, implementation measures, and corresponding management effects should be verified and validated. Verification and validation (V&V) included in scenarios engineering can not only access whether management models, measures, systems, activities throughout the management life cycle are well-functioned to produce high-quality and reliable management decisions but also check whether the management decisions and functions can meet the requirements of both internal demands (e.g., users, customers, etc.) and external environments (e.g., competitors, market, etc.) at the end of management process [39], [40].
Scenarios engineering drives the future management to change from the traditional management process of "POLC" to more complex adaptive "POLC+," where "C+" is represented by command + coordination + control + communication + cognition + cybernetics + · · · . Besides, it will further inject new vitality to the management philosophies that once have been all the range, such as re-engineering enterprises, reengineering management, liberation management, continuity management, continuous organizing, continuous optimizing, etc. Moreover, scenarios engineering can provide powerful technical supports to intelligent, learning, and innovative management that dominated by knowledge works.

B. Models: From AI Foundation Models to Management Foundation Models
According to the research of Stewart [41], knowledge has become the most important element in terms of the percentage of added value of products at the end of the last century. In the past, 80% of the cost was material, and the remaining 20% was knowledge; while the proportion in nowadays has changed to 70% and 30%, respectively. Correspondingly, the significance of knowledge works in management is also constantly highlighted. Aiming to deal with the dynamic environment, increasing complexity and more requirements in a more efficient and economical manner, the research and development of new intelligent tools, namely, management foundation models for managing knowledge production and measuring knowledge works must be accelerated.
The perfection and specialization of the general AI foundation models, such as large language models, big vision models, 3-D models, etc. [42], [43], [44], [45], via integrating the industry-focused and business-oriented foundation models is a feasible way of constructing management foundation models. Due to their inherent nature of "small knowledge, big model, small task, big operation," management foundation models possess the advantages of high generalization so that only very fine-tuning is needed in different management scenarios. Consequently, management foundation models will inevitably develop to be the essential tools for managing knowledge production tasks and make knowledge works be the relatively simpler "physical works" that can be qualitatively measured in smart societies. The widely used robotic process automation (RPA) that replace or assist humans in completing repetitive works can be regarded as the preliminary attempt of this idea.
Management foundation models will bring revolutionary changes to knowledge works and start the journey of management knowledge automation. Based on AI foundation models, the enterprise-level management foundation models should be built as the parallel intelligent "pipeline" for knowledge production, where the proportion of added value of future products is prospected to be comprised of about 5% of materials and 95% of knowledge. Under the parallel management framework, the employment of robots and digital humans will assist and replace humans to produce knowledge and undertake knowledge works, so as to mitigate the excessive requirements of humanity and rationality in the future management. In view of this, we hope that the added value of knowledge pipeline products is approximate to that 80% created by humans, 15% by robots, and 5% by digital humans [23]. This relies on not only the technical route of continuous training and iteration dominated by data and computing powers in AI but also the innovative design of management route in respect to new spaces, new participants, and new processes. The data is used to construct management foundation models via the process of generation, storage, interaction and calculation, and in turn transformed into the precise management knowledge through management foundation models, thus, becomes the dominant intelligence source of the future management.

C. Technologies: CPSS and Metaverses
Management has both engineering complexity and social complexity, in which social complexity involves various participants in different management scenarios and their diversified psychology, goals, behaviors, etc. Management foundation models constructed to measure knowledge works should consider these two types of complexities and bridge the actual and artificial management systems under the parallel management framework, so as to respond to both the regular and long-tail management events and enable data-knowledge-services conversion. The modern management based on cybernetics is to promote the management automation by using knowledge to replace power and money to guide, control and regulate complexity management [46]. This management philosophy is in line with the idea of optimizing the current state, predicting the future state and planning the proper management strategies, but cannot realize the prescriptive intelligence of management that guide knowledge works to strive for the ideal management goals. To this end, the virtual-real interactive CPSS should be used to measure knowledge works.
CPSS is essentially the abstract and scientific definition of metaverse [47]. Metaverses build a virtual world parallel to the real world that can realize the virtualization and digital extension of the real world, based on technologies, including digital twins, extended reality (XR), augmented reality (AR), blockchain, cloud computing, etc. [48], [49]. They closely connect the virtual world with the real world through trust building, digital identity, economic autonomy, and virtual-real integrative applications, thereby reshaping the trust system and management structure in the real world. Metaverses use distributed automated methods driven by smart contracts and intelligent algorithms to promote the significant improvement of management decision-making ability. With CPSS or metaverses as the supporting technologies, the closed-loop feedback from knowledge representation, decision reasoning to scenario-adaptive management can be realized via parallel interaction, thus, providing the effective and practical methods to manage knowledge works [50].
Specifically, humans, robots, and digital humans work together in various management metaverses (ManaMetaVerse) toward specific management problems, collect relevant management knowledge in management multidisciplinary metaverse (ManaMultiVerse), integrate all knowledge in management integration metaverse (ManaInterVerse), generate corresponding management knowledge in management transformation metaverse (ManaTransVerse), and finally form the management scheme with deep intelligence to solve specific problems in management deep metaverse (ManaDeepVerse). It is a breakthrough for measuring knowledge works to build these management metaverses and formulate the knowledge production line that realize the convergence from small data to big data then to deep management intelligence.

IV. NEW MANAGEMENT PARADIGM: SMART ORGANIZATIONS AND INTELLIGENT OPERATIONS
Since humans are the center of all management activities, any management design concerning policies, mechanisms and methods should think about both their individual pursuits and social motivations under the constraints of individual humanity and rationality, to achieve the balance between improving management efficiency and enhancing management satisfaction. Otherwise, the risky behaviors that violate management goals or even undermine organizational development will occur. Taking these into account, the corresponding sociotechnical technology of parallel management must be developed and introduced to eliminate these problems so as to make the management smarter and more sustainable. DAO is just the needed technology that can fully consider humanity and rationality of each member and use collective intelligence to reduce the requirement of individual intelligence. As such, the new management paradigm based on DAOs are proposed for the realization of parallel management.
Based on the distributed blockchain network, DAOs integrate various technologies, such as social computing, crowdsourcing, etc., and make use of smart contracts and democratic voting under the trustless environment to enable stakeholders to effectively participate in the generation, implementation and evaluation process of management decisions [51], [52], [53]. Moreover, DAOs adopt the ecological perspective that pays comprehensive attentions to both hard and soft, humanistic and rational management factors, and make use of both open and closed, internal and external management resources [54], [55], [56]. As such, DAOs can ensure that almost all components that are highly connected with knowledge works play their roles in parallel management. The DAO-based new management paradigm under parallel management framework is named as decentralized management (DeMana) [23]. As follows, the reference model of DAO-based DeMana will be proposed, under which the smart organizations and intelligent operations of DeMana will also be discussed.

A. Reference Model of DAO-Based DeMana
DAOs endow management with ecological sciences and technologies, mainly, including the organizational, coordinational and executional (OCE) technologies [57], and give the comprehensive picture of ecological management. The OCE technologies represented by DAOs can be applied in parallel management to make organizations and enterprises in smart societies achieve "6S" management objectives, namely, safety in the physical world, security in the cyber world, sustainability in the ecology, sensitivity to individual needs, services with good quality, and smartness conforming to human values [59].
With this in mind, the reference model of DAO-based DeMana is proposed in Fig. 3. The OCE technologies of DAOs serve for the management from macro level to meso level and then to micro level, respectively. Organizational technologies primarily fulfill the overall management goals; coordinational technologies mainly support the particular management tasks; and executional technologies focus on solving the specific management problems. Coordinational technologies bridge organizational and executional technologies by converting the goals and information into tasks or works and allocating them to the executive body and equipment. Besides, the upper-level management usually formulates constraints to the lower-level management, while the lower-level management then in turn gives feedback to the upper-level management, which together form closed-loop for DeMana. By virtue of these advantages of DAOs' OCE technologies, Demana will produce and implement trustable, reliable, and usable management decisions in an efficient and effective fashion.
Organizational technologies of DAOs are used to build the organizational form and governance structure of DeMana. The governance structure of community autonomy encourages DAO members to exercise the possessed power to truly represent their own interests in the democratic way, thereby highlighting the role of humanity in management [58]. Meanwhile, the collective decision-making mode in the distributed and decentralized network can break through the restriction of individual rationality, thus, avoiding the adverse consequences caused by the mistakes or evil intentions of someone matters.
Coordinational technologies of DAOs mainly refer to the game-theoretical incentive mechanisms for the distributed collaboration, which drive humans, robots, and digital humans to voluntarily contribute to DeMana with their resources, such as data, knowledge, time, wisdom, etc. [60], [61]. By means of incentives rather than power control, DAOs find a well-suited way to consider and release humanity in the intelligent management. Incentive mechanisms coordinate DAO members to participate in the preferred tasks autonomously thereby transferring the overall goal into task division, and encourage them to jointly achieve the management goal in the way of benign competition and complementary cooperation [62], [63].
Executional technologies of DAOs include not only the blockchain-based smart contracts but also the intelligent algorithms that are often overlooked. They together endow DeMana with powerful executive and learning capabilities, thus, lead to the automated generation and execution of reliable management decisions in real-time.

B. Smart Organizations of DeMana
DeMana carries out the management revolution in the organizational level via organizing members by the common goals, creating the token-based management value system and reshaping the management structure by distributing ownership to all members, thereby capturing the features of smart organizations. Smart organizations of DeMana mainly reflect in smart structures, smart divisions, and smart decisions. 1) Smart Structures: As explained by Ronald H. Coase, companies emerge because they can reduce the costs of production and transaction than individuals via providing high-quality products and services with a competitive cost structure [64]. This is also the reason why people are willing to give up the personal freedom and work together. However, nowadays they have already encountered many fatal problems such as principal-agent problems caused by ownership concentration, high coordinational costs caused by hierarchical structures, etc. [65]. Especially, in the digital age, the organization structure that rely heavily on power and legal jurisdiction are no longer applicable. Instead, more flexible and efficient organizational structures are needed, and DAOs perfectly fit for the desire.
On the one hand, the blockchain-based DAOs lead people around the world to freely aggregate and interact based on common interests, goals and values in the trustless environment, and also share the ownership of DAOs according to their contributions, thus, shaping a smart structure of autonomy. Moreover, this structure can change with the rules and algorithms encoded in smart contracts, making DeMana significantly more flexible and extensible than traditional management structures. On the other hand, DAOs under the parallel management framework integrate the off-blockchain actual organizations and on-blockchain artificial organizations, by introducing the off-blockchian real social relations into the on-blockchain artificial world to guarantee the authenticity and effectiveness of information, and introducing the on-blockchain management experiments into the actual world to assist the optimization and reform of the real management. This closed-loop structure and circular causality of virtual-real interaction makes DeMana truly smart.
2) Smart Divisions: Generally, the division is associated with decision rights allocation, under which who bear risks and acquire rewards correspondingly. Besides of humans, DeMana has also organized robots and digital humans to collaborate together for specific management goals and tasks. The professional, voluntary, and dynamic management divisions of these three types of participants enhance the smartness of DeMana. First, their divisions are enacted according to their capabilities and characteristics. For example, in terms of rationality-driven knowledge works, such as large-scale storage and high-speed computing, robots and digital humans are smarter than humans; while in terms of humanity-driven knowledge works involving complex social relations and social behaviors, humans are smarter than robots and digital humans. Second, their divisions are not mandatory but the natural result of incentive coordination under the established constraints. Everyone chooses to participate in the preferred management tasks that can bring in the desired benefits. Here, benefits are not necessarily the economic incentives represented by tokens or cryptocurrencies but can also be more intrinsic reputation enhancement, trust accumulation, knowledge acquisition, etc. Accordingly, the smart management divisions under which everyone performs his duties and exerts his functions are formed. Third, their divisions change with the organizational structures and management tasks. That is, dynamic specialized divisions, human-machine divisions, and virtual-real divisions are highlighted in DeMana.

3) Smart Decisions:
The decision-making model of DeMana is efficient for both conventional management problems with reference practices and long-tail management problems that require innovative solutions. Because it takes advantages of both DAOs and parallel management to quickly and effectively generate, evaluate, and optimize management decisions. Especially, DAOs unify the ownership and decision-making powers in DeMana, where humans, robots, and digital humans exercise corresponding rights according to their ownership shares, which are obtained at the cost of the valuable contributions. Everyone makes choices voluntarily on his own behalf in the decentralized network, and there is no principal-agent relationship [66]. Besides, DAOs allow the existence of humanity and the resulted self-interested behaviors, and weaken the possible adverse effects through various mechanism design [67]. Democratic voting-based decision mechanism is a good exemplar. It is designed to let the collective interests be represented by the final decision-making results, and the real contributors during the process will be rewarded accordingly [68]. In this way, the self-interested participants can barely obtain high benefits from violating the overall goal of the entire organization. As a result, DeMana can generate more credible and reliable decisions under the premise of fully considering humanity and human values, which is the basic requirement of smart organizations.

C. Intelligent Operations of DeMana
DeMana changes traditional decision-making and implementation methods in the operational level to enhance the intelligent operations of management via the application of smart contracts and intelligent algorithms. The intelligent operations of DeMana are embodied in two aspects: operational automation and operational intelligence.

1) Operational Automation:
DeMana can quickly reach consensus and make management decisions on tasks with specific goals and best practices with the help of smart contracts. Because smart contracts automatically run on blockchain in the programmatic manner and can lead to the quick achievement, dissemination, and implementation of the group consensus [69], [70]. The rules and processes concerning organization, coordination, and execution of DeMana are made into smart contracts after the consensuses being reached, which will then be deployed on blockchain after verification. When the predetermined trigger conditions are satisfied, relevant operations are automatically and accurately executed under the control of codes, so that management requests can be effectively responded to reduce communication costs and improve management efficiency. Meanwhile, smart contracts cannot be tampered arbitrarily, and their merits of transparency, openness and automation make them well-suited to the distributed collaboration management in the trustless environment [71], [72]. This reliable automation brought by smart contracts has significantly improved the operational efficiency of DeMana.
2) Operational Intelligence: Operational automation is the foundation and necessary condition of intelligent operations; while operational intelligence is the key and essence of intelligent operations. Since smart contracts do not have the learning and reasoning abilities and can only passively and regularly perform management operations, they alone cannot realize the evolution from If-Then to What-If [73] and achieve operational intelligence. Instead, intelligent algorithms can actively meet management requirements in strong capabilities of data analysis, learning, and reasoning. In view of this, intelligent algorithms should be introduced into DeMana, which are exactly what should be viewed high by both management researches and applications. Smart contracts and intelligent algorithms can complement to form an organic whole. Intelligent algorithms in DeMana play the crucial roles in various functions, including data, experimentation, decisions, services, etc. They are used to analyze management demands and functions, and encapsulate the formed domain knowledge and decision-making models toward specific management scenarios into smart contracts. As a result, management decisions will be automatically predicted, evaluated, and optimized, rather than simply generated and implemented. Moreover, intelligent algorithms in DeMana are developed under parallel management framework to guide individual and collective behaviors to serve the common goals.

V. NEW MANAGEMENT SOLUTION: MANAGEMENT OPERATING SYSTEMS
The new models and new paradigm under the parallel management framework create conditions for measuring knowledge works and achieving the management sustainability. However, due to the characteristics of multidimensional complexity in management itself, such as complex scenarios, space-time crisscross, numerous participants, etc., the management works of resource allocation, knowledge collaboration, service improvement and efficiency enhancement still lack scalable, safe and efficient interactive solutions. Therefore, we propose to construct MOSs or more general humanoriented operating systems (HOOSs), aiming to enhance the efficiency and effectiveness in managing knowledge works.
The design of MOSs is inspired by the widely-used operating systems, e.g., computer operating systems, robot operating systems [74], [75]. MoSs are essentially the blockchain-based human-machine interactive platforms comprised of a collection of program modules that control the running and operations of various hardware and software resources, e.g., network, data, models, algorithms, knowledge, incentive, etc. (as shown in Fig. 4). MOSs mainly include decentralized applications (DApps) and kernel, where DApps request the relevant business to kernel and get response from it. By supporting dynamic configuration of plugins, MOSs decouple the business logic of DApps from the underlying blockchain technologies and also provide developers with friendly API interfaces. Smart contracts are used to automatically perform all actions in MOSs. In this way, diverse management activities of resource allocation, process control, knowledge collaboration, decision optimization, and service development are provided by MOSs. Besides of actual MOSs, artificial MOSs parallel to them will also be built. Artificial MOSs conduct management experiments and explorations that cannot be carried out in actual MOSs, including scenarios engineering, data expansion, knowledge specification, state prediction, strategy evaluation, personnel training, and decision optimization. Parallel execution between these two MOSs will promote the realization of 6S goals in smart societies.
From the perspective of managers, MOSs reasonably organize and manage the workflow and maintain the management environment, supporting the secure exchange, efficient collaboration and reasonable configuration of data, knowledge, business, and other resources; from the perspective of users, MOSs eliminate the differences of various infrastructures and provide safe, convenient, and efficient functional services. MOSs aim to re-engineer the flow of knowledge works and reorganize the logic of management business, realize reliable production and safe sharing of data and knowledge, conduct unified control of management architectures, models and technologies, and, thus, form a benign management ecology of sustainable development. To accomplish these goals and realize management intelligence, the features of simple intelligence, provable security, flexible scalability, and ecological harmony should be highlighted in the design and construction of MOSs.
Simple Intelligence: MOSs provide users with general management software in the form of interfaces to assist the organization and completion of knowledge works. When faced with the large numbers of intelligent information processing equipment and technologies with complex associations, simple but effective design of MOSs are particularly important. Simple intelligence does not mean that MOSs can only provide primary intelligent programs or functions. Instead, it is hoped that advanced and complex blockchainbased intelligent functions can be presented to users in an easy-to-understand and easy-to-use way [76], in favor that they can manage knowledge works more conveniently and efficiently.
Provable Security: MOSs ensure the comprehensive security from all aspects of management, including data basis, consensus process, and system state. The traditional security solutions are to consider system security as a software function added to the existing systems or solved via additional hardware, which are far from enough for the security required by smart organizations and intelligent operations of the future management. Unlike this, MOSs themselves are the intelligent secure systems having a mathematically provable secure blockchain-based kernel and unbreakable hardware [77], which together provide security to manage knowledge works.
Flexible Scalability: MOSs are built on the cloud-based hardware infrastructure and adopt the modular strategy to organize various management functions. Under the unified framework, foundation models, languages and platforms, the structured, and customized and reusable plugins are designed and built for MOSs. Then, the functional modules oriented to specific management scenarios and business logic are formulated by means of the agile configuration and coordination of different plugins. They further integrate to be a flexible and dynamic adaptive management architecture [78].
Ecological Harmony: MOSs take humans and their knowledge works as the core, connect a large number of management equipment, data, software, and support businesses, products and services, thus, forming an open development ecology. Therefore, the construction and running of MOSs are not only deeply integrated with the internal and external management environment of the real world but also with the management environment of the virtual world. Moreover, MOSs take into account of management spaces, participants, resources objectives and behaviors, and harmonize them to form a sustainable development balance [79].

VI. CONCLUSION
To measure knowledge works and manage the knowledge production in smart societies, this article first proposes parallel management framework by comprehensively integrating ACP theories and the supporting intelligent technologies, which can realize the descriptive, predictive, and prescriptive management intelligence by means of the virtual-real interaction, closed-loop feedback and parallel execution of the physicspace, cyberspace, and social space. Then, the management foundation models are built based on AI foundation models, which use scenarios engineering to construct trustworthy management scenarios and CPSS to support the knowledge production. Further, the new management paradigm, namely, DeMana and its reference model is put forward, which relies on the ecological OCE technologies of DAOs to realize smart organizations and intelligent operations of management. Finally, MOSs are designed to provide secure and scalable solutions for managing knowledge works. The proposed new framework, model, paradigm, and solution together promote the future management to achieve 6S goals.
Although we have proposed to form the ecological mode of organization, coordination and execution by humans, robots, and digital humans under parallel management framework, there still lack the concrete technical supports, such as heterogeneous virtual population generation, character and cognitive modeling, natural language processing and incentive designing of digital humans. Meanwhile, when more virtual-real interactive decision-making spaces and decisionmakers are included, the complexity of management decisions may exceed the current imagination, which drives in-depth studies on the corresponding decision-making methods, technologies and models. Moreover, we believe that management foundation models and MOSs are the inevitable trend of future management, which urge us to explore efficient management foundation models and construct intelligent MOSs.