複雜性科學視野下的中國軍隊指揮控制變革趨勢
現代英語：
Hu Xiaofeng
Introduction
In recent years, with the evolution of war forms, the epistemology under the traditional scientific system has become increasingly difficult to meet the needs of the development of war practice. Complexity science has provided new possibilities for understanding war and guiding war practice. In the field of command and control, complexity can be used as a weapon to make the opponent’s decision-making more complicated, while the opponent can easily deal with it. The key to achieving this effect is to break the original traditional decision-making method and reshape the opponent’s decision-making process.
Understanding the Principle of Complexity in War
To discuss “how complexity becomes a weapon”, we first need to understand where the complexity of war comes from and the nature and characteristics of complex systems.
Traditional thinking methods believe that the world is deterministic and people are rational. Humans can use methods such as “decomposition + cognition + synthesis” to achieve a comprehensive understanding of the world and a precise grasp of the nature and operating laws of all things. However, there are still many complex systems in the world. These systems have holistic properties. Systems such as the human body, society, economy, and war all belong to this category. Complex systems have variable structures and are characterized by adaptability, uncertainty, emergence, and nonlinearity. Moreover, the results are not repeated and difficult to predict. Society, economy, cities, including intelligent systems, these systems related to people have these characteristics, in fact, they are all typical complex systems.
War is a typical complex system. People often say “randomness in war”, “unrepeatable results of war”, “fog in war”, “no constant situation of soldiers, no constant shape of water”, etc., in fact, they are all describing the characteristics of war complexity.
The root cause of war or social complexity lies in people’s “cognition”. People are not chaotic particles without thoughts, nor are they low-level creatures with only simple life logic. Therefore, the use of simple physical statistical methods will produce great deviations. Moreover, human behavior is not just “pure rationality” at work. Emotional factors such as personality, emotions, and attitudes may affect changes in human behavior.
Recognizing that war is a complex system is a huge progress. Only in this way can we better understand the phenomenon of war complexity and find ways and means to solve problems. Instead of simply adopting reductionist methods or simply transplanting physical methods to study war.
How to create complexity
The earliest idea of making complexity a weapon was the US military’s “decision-making center war”. It advocates that with the support of advanced technologies such as artificial intelligence, the upgrading and transformation of combat platforms and distributed deployment can achieve diversified tactics, while ensuring its own tactical selection advantages, it can impose high complexity on the opponent to interfere with its command and decision-making capabilities, and occupy an overwhelming advantage in the new dimension.
In future wars, due to the increasingly large combat system, the interweaving of people and machines in decision-making at all levels, especially the widespread use of intelligent combat platforms, the battlefield management problem has become unprecedentedly prominent. If there are defects in the opponent’s information system, it may be subject to the complexity of its own system. The US military believes that complexity as a weapon is to make the opponent feel complex in one direction, while it can easily deal with it. The key is to break the original traditional decision-making method and reshape the opponent’s decision-making process.
Traditional decision-making is only a change in decision-making parameters, not a change in decision-making process, so the decision complexity is constant. If the decision is still made according to inherent factors and processes, only the size of the variable is adjusted, and the original causal relationship is not changed. For example, the self-side adopts camouflage, reinforcement and other means to enhance the survivability of the base and force the opponent to change the original decision. But doing so does not increase the complexity of the enemy’s decision-making, because the causal relationship has not changed. If complexity methods are introduced, the opponent’s decision-making process can be reshaped, forcing the opponent to introduce new parameters, resulting in increased decision-making complexity. For example, the camouflage effect of one’s own side exceeds the opponent’s existing reconnaissance capability, forcing it to find new reconnaissance and positioning methods, thereby prompting it to change its decision-making process and make the decision more complicated. So, how to create complexity? It is generally believed that there are four main methods.
Limit intelligence capabilities. By creating uncertainty to enhance complexity, reduce the opponent’s situational awareness, so that the opponent can only take action with the support of limited information. Specific methods include: giving incomplete information or false information, which is what we often call hiding the truth and showing the false; creating deep uncertainty, that is, creating “black swan” events, putting the opponent in an “unknown unknown” state. For example, electronic countermeasures make it impossible for the opponent to understand the system principles, resulting in an inability to understand the situation. The most typical case is the “unknown air situation” in the past. Although it can be seen from the radar signal, it is unknown why such a signal is generated, which makes the opponent always confused.
Weaken response capabilities. By using adaptive characteristics to create complexity, the opponent’s action response capability is weakened. It can also be understood that changing the environment faster creates complexity, forcing the opponent to reduce its adaptability and hindering its timely feedback. Combat confrontation can be seen as who can adapt to changes in the opponent, battlefield and environment faster. The adaptability depends on the command and control system structure, process and decision-making method, as well as the reaction speed of the troops. Without this ability, it is often passive. Adopting flexible command methods, making adequate plans, and having distributed decision-making mechanisms can enhance the adaptability of the troops. The traditional centralized control method will be difficult to adapt to the needs of future wars, and the dispatch-type command and control mode with resource distribution and service distribution similar to online car-hailing may be better adapted to the changing battlefield environment and respond faster.
Create border chaos. Create chaos and complexity by crossing boundaries to form new emergent effects. Between military services and between different support systems, etc., can all be regarded as boundaries. In future operations, cross-service and cross-system command boundaries often become system vulnerabilities, forming more complex network cascade effects. Therefore, how to create complexity across organizational boundaries will become the key to future operations. For example, creating all kinds of chaos between different combat domains, different troops, and different systems, while ensuring the orderly operation of one’s own system, making the opponent’s system entangled and inefficient, thus causing its system to be divided or collapsed.
Promote critical point conversion. Push the opponent to the critical point and create complexity, resulting in nonlinear conversion. Small changes in complex systems may be amplified and transmitted step by step, and emerge with effects that seriously affect the war situation, and this phenomenon often occurs at the critical point. The most common practices are “overload” and “promote collapse”. The so-called overload is to force the opponent to be in an overloaded state frequently, resulting in a significant reduction in its performance. For example, the command information system may perform very well at ordinary times, but once it is under high overload during wartime, its performance will decline significantly and it will often be in a low-efficiency state. The so-called collapse is to force the enemy system to collapse at the critical point. For example, the commander collapses under high pressure during wartime and makes a wrong judgment, even if he may only be at a slight disadvantage. In this case, the firmness of the commander’s will often determines the final result.
“Fight complexity with complexity”
Almost all changes in war can be seen as cognitive upgrades and complications. Clausewitz believed that “uncertainty” and “fear” are the two keys to studying war, and the commonality between the two is actually cognition.
The key to “decision-centered warfare” is to accelerate cognition to adapt to complexity. To deal with complex weapons, we need to “fight complexity with complexity” and start with basic work. First, we need to understand the principles of war complexity and actively create complexity to seize the initiative on the battlefield. For ourselves, we need to manage our own complexity; for our opponents, we need to impose more complexity on them. Secondly, we need to understand the changes in the mechanism of war, which requires us to focus on the key point of “command and control”. The most important of these is the perception, control and management of complex battlefields, as well as the understanding, modeling and simulation of intelligent cognition.
As we all know, command and control has two sides. Its “rational” side is the standardization of command decisions, which is reflected in the command mechanism, combat process, regulations, plans and planning. That is, “when you know how to do it”, you can use efficient scientific methods to make it, such as operational planning. The “emotional” side is reflected in the creative content of command decisions, that is, the art of command. The art of command depends on the inspiration, intuition, knowledge and experience, and even personality of the commander. That is, when “I don’t know what to do”, I can find a way to solve the problem, such as trial and error. In fact, the former is only the result of intelligence, and the latter determines the generation of intelligence. Therefore, from a rational point of view, traditional systems hope that everything that will happen in the future is known in advance; from a perceptual point of view, intelligent systems hope that surprises will emerge in the future, that is, innovation. Therefore, it can be considered that command is art and control is science; command is “intelligence” and control is “skill”, which is the difference between “strategy” and “technology”. How to solve these two problems is the trend and direction of future changes in the field of command and control.
The first trend: control management is moved down, and intelligent decision-making is continuously mechanized. Under the conditions of future informationized and intelligentized warfare, the independent operation of unmanned systems, the adaptive organization of military organizations, and the autonomous combat methods based on big data, cloud computing and network information systems have put forward higher requirements for the command and control mode. To adapt to this change, it is necessary to give combat units more autonomy and shorten the chain of command; introduce intelligent decision-making assistance and analysis tools, and use people “on the loop” or “outside the loop” to effectively accelerate the decision-making process and improve decision-making efficiency.
The second trend: command decisions are moved up and become more complex with intelligent assistance. From a strategic perspective, first, we should focus on cognitive confrontation, fully understand complex system problems such as “the more complex the system is, the more complex it is”, “yesterday’s rules are different from today’s”, and “the future is unpredictable”, and find ways to deal with them; second, we should confront complexity with complexity. “Strategic simplicity” cannot replace “the complexity of battlefield and tactical control”, and we cannot ostrich-like “respond to changes with the same”. Third, to create complex asymmetry, combat personnel need to use “command art + decision-making intelligent tools”. The stronger the artistry of decision-making, the higher the complexity. To fight against artificial intelligence, we need to rely on artificial intelligence, but we cannot rely on artificial intelligence alone.
The third trend: build a combat network and command and control network system around intelligent machines. First, we need to redefine the platform type and no longer simply distinguish between combat platforms of different space types such as aircraft and submarines. Second, we need to change the traditional combat force use mode and adopt an “order-based” combat mode. Focusing on combat resources and services, we provide battlefield links, command, management and scheduling through cloud platforms. The third is to improve the adaptability of complex systems, so that the command and control and combat systems can bear high overload situations, and cannot perform “excellently” in simple situations, but seriously decline in complex situations, leading to complex disasters.
The fourth trend: Human-machine hybrid command and control will become the mainstream. In the future, with the support of big data, combat cloud, machine assistance and shared combat maps, artificial intelligence will form a combat ecosystem with humans. It should be noted that the decision-making application of artificial intelligence is not to replace “people”, but to make good use of “people”, and achieve better results through human-machine collaboration. Specifically, one is to combine human creativity with the “tirelessness” of machines based on hybrid intelligence, so as to give full play to their respective strengths and avoid the dilemma that decision-making intelligence technology has been difficult to break through for a long time. The second is to solve the problem of human-machine hybrid. The inconsistency of cognitive space, the difficulty of forming a collaborative system, and the distrust of people in machines all need to be solved, which may be the focus of future command and control system design.
現代國語：
胡曉峰
引言
近年來，隨著戰爭形態的演變，傳統科學體系下的知識論越來越難以滿足戰爭實踐發展的需要，複雜性科學為認識戰爭、指導戰爭實踐提供了新的可能性。在指揮控制領域，複雜性可以作為一種武器，使對手決策變得更加複雜，而自己則可以輕鬆應對。要達到這效果的關鍵在於，打破原有傳統決策的方式，重塑對手的決策流程。
認識戰爭複雜原理
討論“複雜性怎樣成為武器”，我們首先需要認識一下，戰爭的複雜性由何而來，以及複雜系統的性質和特徵。
傳統思維方法認為，世界是確定性的、人是理性的，人類可以採用「分解+認識+綜合」等方法，實現對世界的全面認識、對萬事萬物的性質和運作規律的精準把握。但世界上還存在著許多複雜系統，這些系統存在著整體性質，像是人體、社會、經濟、戰爭等，都屬於這一類。複雜系統結構可變，具有適應性、不確定性、湧現性、非線性等特點，且結果不重複，也難以預測。社會、經濟、城市包括智慧系統，這些與人有關的系統都有這些特點，其實它們都是典型的複雜系統。
戰爭就是典型的複雜系統。人們常說的「戰爭中的偶然性」「戰爭結果的不可重複性」「戰爭中的迷霧」「兵無常勢，水無常形」等等，其實都是在描述戰爭複雜性的特徵。
戰爭或社會複雜性產生的根源，在於人的「認知」。人不是雜亂無章、沒有思想的粒子，也不是只有簡單生命邏輯的低等生物，因而採用簡單的物理統計方法就會產生很大的偏差。而且，人的行為也不會只是「純理性」在起作用，性格、情緒、態度等感性因素都有可能影響人類行為的改變。
認識到戰爭是複雜系統，是一個巨大進步。唯有如此，才能更能理解戰爭複雜性現象，找到解決問題的方法和途徑。而不是簡單地採用還原論方法，或簡單地移植物理學方法來研究戰爭。
如何製造複雜性
讓複雜性成為武器，最早是美軍「決策中心戰」的設想。其主張在人工智慧等先進技術的支撐下，透過對作戰平台的升級改造以及分散式部署實現多樣化戰術，在保障自身戰術選擇優勢的同時，向對手施加高複雜度，以乾擾其指揮決策能力，在新維度上佔據壓倒性優勢。
未來戰爭，由於作戰體系越來越龐大，各級決策中人與機器混雜交織，尤其是智慧化作戰平台的廣泛運用，導致戰場管理問題空前突出。如若對手資訊系統上有缺陷，可能將受制於自己體系的複雜性。美軍認為，複雜性作為武器，就是要讓對手單向感到複雜，而自己則可以輕鬆應付。其關鍵在於打破原有傳統決策的方式，重塑對手的決策流程。
傳統決策只是決策參量的改變，而非決策流程的改變，所以決策複雜度恆定。如若仍依固有因素和流程決策，調整的只是變數的大小，並未改變原先的因果關係。例如，己方透過採取偽裝、加強等手段，增強基地的生存能力，迫使對手改變原有決策。但這樣做並沒有增加敵方的決策複雜度，因為因果關係並沒有改變。如果引入複雜性方法，則可以重塑對手的決策流程，逼迫對手引入新參量，導致決策複雜性增加。例如，使己方偽裝效果超過對手現有的偵察能力，迫使其必須尋找新的偵察定位方法，促使其改變決策流程，使決策變得更加複雜。那麼，如何製造複雜性呢？一般認為，主要有以下四種方法。
限制情報能力。透過製造不確定性以增強複雜性，降低對手的態勢感知能力，使得對手只能在有限資訊的支撐下實施行動。具體的方法包括：給予不完整的資訊或虛假訊息，也就是我們常說的隱真示假；製造深度不確定性，也就是製造「黑天鵝」事件，讓對手處於「未知的未知」狀態。例如，電子對抗戰法使對手無法理解系統原理，導致無法理解態勢。最典型的案例莫過於過去的“不明空情”，雖然可以從雷達信號上看到，但不知道為何會產生出這種信號，使對手始終處於迷茫狀態。
削弱反應能力。透過利用適應性特徵來製造複雜性，削弱對手的行動反應能力。也可以理解為，更快變換環境製造複雜性，迫使對手降低適應能力，並阻礙其及時回饋。作戰對抗可以看作是敵我雙方誰能更快地適應對手、戰場和環境的變化，適應能力取決於指控系統結構、流程和決策方式，以及部隊的反應速度。沒有這方面的能力，往往就會陷入被動。採用靈活的指揮方式，做好充分預案，具有分散式決策機制等，都可以增強部隊的適應性。傳統意義上的集中控制方式將難以適應未來戰爭的需要，而類似於叫車那樣的資源分佈、服務分佈的派單式指揮控制模式，或許才能更好地適應多變的戰場環境，並能做出更快的反應。
製造交界混亂。透過跨越邊界製造混亂、創造複雜性，形成新的湧現效果。軍兵種之間、不同保障系統之間等，都可以看成是邊界。未來作戰中跨軍種、跨體系的指揮邊界往往成為體系脆弱點，形成更複雜的網路級聯效應。因而如何在跨組織邊界製造複雜性，將成為未來作戰的關鍵。例如在不同作戰域、不同部隊、不同系統之間製造各種混亂，在確保己方體系運作有序的同時，使對手體系相互纏繞低效，從而導致其體系分割或坍塌。
促進臨界點轉換。將對手逼到臨界點並製造複雜性，造成非線性的轉換。複雜系統中的微小變化，都可能經過逐級放大傳導，湧現出嚴重影響戰局的效果，而這個現象經常出現在臨界點處。最常見的做法是「超負荷」「促進崩潰」兩種。所謂超負荷，就是促使對手經常處於超負荷狀態，導致其性能大幅降低。例如，指揮資訊系統平時可能表現很優秀，但在戰時一旦處於高過載情況下，表現卻會大幅下滑，常常處於低效能狀態。所謂促崩潰，就是促使敵方系統在臨界處崩潰。例如指揮在戰時高壓力下崩潰作出錯誤判斷，就算其可能只是處於微小劣勢之下。在這種情況下，指揮官意志的堅定程度，往往決定了最終的結果。
“以複雜對抗複雜”
戰爭中幾乎所有的變化，都可以看成是認知的升級和複雜化。克勞塞維茲就認為，「不確定性」和「恐懼」是研究戰爭的兩個關鍵，這兩者的共通點其實就是認知。
「決策中心戰」的關鍵就在於加速認知，以適應複雜性。應對複雜性武器需要“以複雜對抗複雜”，從基礎工作做起。首先，需要理解戰爭複雜性原理，主動製造複雜性來掌握戰場主動權。對自己來說，需要管理自身的複雜性；對對手來說，需要對對手施加更多的複雜性。其次，需要了解戰爭機理發生的改變，這就需要瞄準「指揮與控制」這個重點。其中最重要的是對複雜戰場的感知、控制和管理，以及對智慧認知的理解、建模和模擬。
眾所周知，指揮控制具有兩面性。它「理性」的一面，就是指揮決策的規範化，體現在指揮機制、作戰流程、條令條例、計畫規劃等內容。也就是「知道怎麼做時」可以用高效率的科學方法做出來，例如運籌規劃。 「感性」的一面則體現在指揮決策的創意內容，也就是指揮藝術。指揮藝術取決於指揮人員的靈感直覺、知識經驗，甚至是性格。也就是在「不知道怎麼做時」能找到解決問題的方法，例如試誤選擇。事實上，前者只是智能結果，後者才決定智能的產生。所以，從理性角度來看，傳統系統希望未來發生的一切都是事先知道的；而從感性角度來看，智慧系統卻希望未來能夠湧現出驚喜，也就是創新。所以可以認為，指揮是藝術，控制是科學；指揮是“智能”，而控制是“技能”，這也就是“謀略”和“技術”的區別。而如何解決好這兩個問題，是未來在指揮控制領域變革的趨勢與方向。
第一個趨勢：控制管理下移，並將智慧決策不斷機器化。在未來資訊化智慧化戰爭條件下，無人系統獨立運作、軍隊組織自適應編組以及基於大數據、雲端運算和網路資訊系統的自主作戰方式等，對指揮控制模式提出了更高的要求。適應這種變化，需要賦予作戰單元更多自主權，縮短指揮鏈；引入智慧化輔助決策與分析工具，以人在「迴路」或「迴路外」等方式，切實加速決策過程，提高決策效率等。
第二個趨勢：指揮決策上移，在智慧輔助下更為複雜。從謀略角度講，一是要以認知對抗為中心，充分理解「複雜系統越認識越複雜」「昨天規律與今天不同」「未來不可預測」等複雜系統問題，找到應對之道；二是以複雜對抗複雜。 “戰略的簡潔性”並不能取代“戰場和戰術控制的複雜性”，不能鴕鳥式的“以不變應萬變”。第三是創造複雜的非對稱性，需要作戰人員藉助「指揮藝術+決策智慧工具」。決策的藝術性越強，複雜度越高，對抗人工智慧需要依賴人工智慧，但不能只靠人工智慧。
第三個趨勢：圍繞智慧機器，建構作戰網和指控網體系。一是需要重新定義平台類型，不再簡單區分飛機、潛水艇等不同空間類型的作戰平台。二是需要改變傳統作戰力量使用模式，採取「訂單式」作戰模式，圍繞作戰資源與服務，透過雲端平台提供戰場連結、指揮、管理與調度。第三是提高複雜系統適應能力，指揮控制與作戰體系能夠承擔高過載情況，不能遇到簡單情況表現“優秀”，但在復雜情況下卻嚴重下滑，導致複雜性災變。
第四個趨勢：人機混合的指揮控制將成為主流。未來的指揮控制，在大數據、作戰雲、機器輔助和共用作戰圖的支援下，人工智慧將和人一起形成作戰生態系統。要注意的是，人工智慧的決策運用，不是要取代“人”，而是要用好“人”，人機協同，實現更好的效果。具體來說，一是基於混合智能，將人的創造性與機器的「不知疲倦」結合起來，既發揮各自的長處，又規避決策智能技術長時期難以突破的窘境。二是解決人機混合的難題。認知空間的不一致、協同體系的難形成、人對機器的不信任等，都需要解決，或許是未來指控係統設計的重點。