U.S. Strategic Competition with China

A RAND Research Primer

by Timothy R. Heath

更多

A rapid unraveling of the U.S.-China relationship—which had been widely viewed as stable and mutually profitable despite long-standing disputes—has unsettled global politics. Although both capitals appear committed to peacefully resolving their differences, the intensifying acrimony and distrust have raised fears among many observers that the two countries could be headed toward confrontation.

The author reviews past RAND Corporation studies on relevant topics surrounding this relationship, primarily from the past five years and extending to the middle of 2020. In this examination of more than 60 RAND reports, the author discusses various issues pertaining to the competition, including China's strategic goals and priorities, the policies and measures through which China attempts to fulfill these goals, how China's actions affect U.S. strategic interests, and what additional steps might further protect U.S. interests. This review also encompasses studies that take a broader view of the competitive international environment and consider how the competition might differ from U.S. strategic competition with Russia. The author excluded from this review tactical-level studies that offer detailed comparisons of U.S. and Chinese military capabilities or examine the value of specific assets or systems under certain warfighting conditions.

This review first highlights major findings across the RAND studies, then discusses key themes that RAND researchers have found particularly relevant in the context of U.S.-China strategic competition. A concluding section outlines topics for additional research that could further inform the work of strategy and policy development.

收起

Understanding Influence in the Strategic Competition with China

by Michael J. Mazarr, Bryan Frederick, John J. Drennan, Emily Ellinger, Kelly Elizabeth Eusebi, Bryan Rooney, Andrew Stravers, Emily Yoder

Related Topics:

更多

China,

Intelligence Analysis,

International Diplomacy,

International Economic Relations,

Southeast Asia

Over the past two decades, China's role in the geopolitical landscape has grown, particularly as a result of the country's rising economic and military power. Thus, U.S. leaders now view China as a strategic competitor—one that seeks to upend the post–World War II liberal international order. One of China's strategies in that competition is to seek influence in countries around the world. In this report, the authors assess China's ability to use various mechanisms of influence to shape the policies and behavior of 20 countries, as well as the lessons that these examples offer for the United States' strategic competition with China. With this study, the authors aim to produce a transferable framework (comprising inputs, intervening factors, and outputs) and other tools of analysis that can provide reliable means of assessing bilateral influence relationships in other cases.

Among the study's chief findings is that China's burgeoning economic power, above and beyond any other considerations, is the foundation for its influence. Furthermore, Beijing's ability to manipulate local political, economic, and social events to its benefit is an important factor in its influence efforts. If China's mammoth economic magnet is the gravitational center of its influence, its ability to reach into other countries and effectively manipulate perceptions and events is the predominant tool. Nevertheless, success in the competition for influence is as much about how the United States responds to current challenges as it is about anything China does or does not do.

Key Findings

The most effective source of influence in the various cases and data sources that the authors investigated was the weight of attraction of China's economy.

The most easily measurable components of influence—the inputs (e.g., trade, foreign direct investment, number of visits by senior officials, presence of an influence country's media)—are imperfect predictors of outcomes.

As a result, inputs are not necessarily the best indicators of potential influence and go only so far in defining the scope and severity of the influence challenge.

The second-most-effective source of influence was China's targeted, often clandestine outreach—sometimes with financial incentives—to specific leaders, elites, and opinion-shapers in targeted countries.

The contest for influence is primarily, indeed dominantly, nonmilitary in character.

China faces a dilemma in exercising influence: The more strongly China tries to use its clout to force outcomes in other countries, the greater the backlash it foments and the more those countries reject its influence.

The most generalized geopolitical reaction to growing Chinese power is hedging.

Despite that predominant trend of hedging, the general geopolitical alignment of the focus countries that we assessed for this study remains highly positive for the United States.

The emerging competition for influence is not strictly government to government.

Recommendations

For countries that are the target of Chinese influence, the United States should continue to offer the credible promise of assistance when they stand up to Chinese coercion.

The U.S. government needs a comprehensive approach to understanding, tracking, and responding to China's targeted influence programs aimed at specific leaders or elites.

If the United States is not willing to simply accept the unqualified effect of China's economic growth, it must do much more to develop investment tools capable of offering countries an alternative to Chinese money—especially in infrastructure.

The United States should support independent research, journalism, and transparency initiatives focused on China's activities in key countries.

The United States should develop an influence strategy with a broader network (beyond journalists) to rally nongovernmental sources of opinion and action in targeted countries.

Table of Contents

Chapter One

Introduction

Chapter Two

Defining Power and Influence

Chapter Three

Framework for Assessing Chinese Influence

Chapter Four

Measuring Chinese Influence: Inputs

Chapter Five

Measuring Chinese Influence: Intervening Factors

Chapter Six

Measuring Chinese Influence: Outputs

Chapter Seven

Case Studies of Chinese Influence-Seeking

Chapter Eight

Historical Influence Case: Finland in the Cold War

Chapter Nine

Overall Findings and Implications

收起

Air Dominance Through Machine Learning

A Preliminary Exploration of Artificial Intelligence–Assisted Mission Planning

by Li Ang Zhang, Jia Xu, Dara Gold, Jeff Hagen, Ajay K. Kochhar, Andrew J. Lohn, Osonde A. Osoba

更多

U.S. air superiority, a cornerstone of U.S. deterrence efforts, is being challenged by competitors—most notably, China. The spread of machine learning (ML) is only enhancing that threat. One potential approach to combat this challenge is to more effectively use automation to enable new approaches to mission planning.

The authors of this report demonstrate a prototype of a proof-of-concept artificial intelligence (AI) system to help develop and evaluate new concepts of operations for the air domain. The prototype platform integrates open-source deep learning frameworks, contemporary algorithms, and the Advanced Framework for Simulation, Integration, and Modeling—a U.S. Department of Defense–standard combat simulation tool. The goal is to exploit AI systems' ability to learn through replay at scale, generalize from experience, and improve over repetitions to accelerate and enrich operational concept development.

In this report, the authors discuss collaborative behavior orchestrated by AI agents in highly simplified versions of suppression of enemy air defenses missions. The initial findings highlight both the potential of reinforcement learning (RL) to tackle complex, collaborative air mission planning problems, and some significant challenges facing this approach.

Key Findings

RL can tackle complex planning problems but still has limitations, and there are still challenges to this approach

Pure RL algorithms can be inefficient and prone to learning collapse.

Proximal policy optimization is a recent step in the right direction for addressing the learning collapse issue: It has built-in constraints preventing the network parameters from changing too much in each iteration.

ML agents are capable of learning cooperative strategies. In simulations, the strike aircraft synergized with jammer or decoy effects on a SAM.

Trained algorithms should be able to deal with changes in mission parameters (number and locations of assets) fairly easily.

Few real-world data exist on successful and unsuccessful missions. Compared with the volumes of data used to train contemporary ML systems, very few real missions have been flown against air defenses, and virtually all of them were successful.

For analyses involving the use of large simulations in place of large datasets, the required computational burden will continue to be a significant challenge. The scaling of computational power and time required to train realistic sets of capabilities (dozens of platforms) against realistic threats (dozens of SAMs) remains unclear.

Developing trust in AI algorithms will require more-exhaustive testing and fundamental advances in algorithm verifiability, and safety and boundary assurances.

Recommendations

Future work on automated mission planning should focus on developing robust multiagent algorithms. Reward functions in RL problems can drastically change AI behavior in often unexpected ways. Care must be taken in designing such functions to accurately capture risk and intent.

Although simulation environments are crucial in data-scarce problems, simulations should be tuned to balance speed (lower computational requirements) versus accuracy (real-world transferability).

Table of Contents

Chapter One

Introduction

Chapter Two

One-Dimensional Problem

Chapter Three

Two-Dimensional Problem

Chapter Four

Computational Infrastructure

Chapter Five

Conclusions

Appendix A

2-D Problem State Vector Normalization

Appendix B

Containerization and ML Infrastructure

Appendix C

Managing Agent-Simulation Interaction in the 2-D Problem

Appendix D

Overview of Learning Algorithms

2020年5月31日，美国兰德公司发布《通过机器学习实现空中优势：对人工智能辅助任务规划的初步探索》（Air Dominance Through Machine Learning：A Preliminary Exploration of Artificial Intelligence–Assisted Mission Planning）研究报告，该研究为人工智能原型系统在空战环境中开发和评估新型作战概念的潜力提供了证据支撑。

研究团队测试了几种学习技术和算法来训练能够在模拟环境中进行空战规划的智能代理，目标是利用人工智能系统的能力大规模地反复模拟和持续改进，从而加速并丰富作战概念的发展。报告指出，人工智能任务规划工具相比现有的人工或自动规划技术将具有极大的速度优势。

该研究最大的目的是在简化的仿真模拟环境中验证人工智能系统进行空战任务规划进而创新战法的应用潜力，值得关注的亮点和结论有：1）假想的场景是给定一组配置不同传感器、武器、诱饵和电子战载荷的无人机对孤立防空体系进行进攻，场景的选取贴合蜂群、“马赛克战”等的基本构想；2）将开源深度学习框架与国防部标准作战模拟工具“仿真、集成与建模高级框架”（AFSIM）集成，分别测试了生成对抗网络（GAN）、Q学习、异步优势动作评价（A3C）、近端策略优化（PPO）等当前最新的开源算法，结果显示只有PPO算法可以在一组变化的复杂场景下满足任务规划需求；3）人工智能算法在任务规划中展现了一定优势和潜力，倘若在算法、训练和部署中进一步深化研究，有望在速度上相比现有的人工或自动任务规划带来巨大优势；4）强化学习是在数据资源稀缺情况下利用人工智能实现作战概念和战术创新最有力的工具，回报函数的设置是决定人工智能演进策略的关键所在，是作战智慧的核心涌现；5）在真实场景和关键任务中使用基于统计学习方法的人工智能系统必须经过充分的试验、测试和认证。因此必须重视和加快航空可信人工智能框架方面的工作，其重要性不亚于其他航空人工智能应用探索工作。

收起

Fundamental Research Policy for the Digital Battlefield

by Leland Joe, Phillip M. Feldman

The Department of Defense is looking to commercial information technologies to meet its needs for digitization equipment. The commercial marketplace has shown responsiveness and agility in meeting the growing civilian demands for robust, reliable, and ubiquitous communications. Many of these technologies are of direct use or can be leveraged to develop systems for the military. This study examines the ability of the commercial marketplace to meet the future needs of the Army, and it identifies research areas for Army investment. The study focuses on identifying fundamental communications network characteristics (physical topology, operating environment, user needs) that uniquely define the Army's communications problem and are not being addressed by commercially driven research. The authors develop a framework that links the Army's future operational capabilities to system design tradeoffs. This framework is then used to examine how well commercial systems can meet Army needs. The authors find that commercial wireless systems will not meet the Army's future needs, and the Army needs to trade off requirements with future investments in research and Army-unique systems. Specific recommendations for Army investment in specific technologies are provided in a companion report: Phillip M. Feldman, Emerging Commercial Mobile Wireless Technology and Standards: Suitable for The Army?

On Distributed Communications

Summary Overview

by Paul Baran

更多

One in a series of eleven Memoranda detailing the Distributed Adaptive Message Block Network, a proposed digital data communications system based on a distributed network concept. It summarizes the development of the system proposal. The more salient features of the system (given in greater detail in the other Memoranda in the series) are outlined, giving the casual reader an overview of the network. A reading of the introductory number of the series (RM-3420) and this one should suffice for a general understanding of it. Additionally, the advantages and disadvantages of the system are given and will greatly facilitate critical evaluation of the system's feasibility. (See also RM-3097, RM-3103, RM-3420, RM-3578, RM-3638, RM-3762 through RM-3766.)

收起

On Distributed Communications

X. Cost Estimate

by Paul Baran

更多

One in a series of eleven Memoranda detailing the Distributed Adaptive Message Block Network, a proposed digital data communications system based on a distributed network concept. It presents the equipment cost estimates for a proposed system based on an arbitrary network configuration consisting of 400 switching nodes servicing 100,000 simultaneous users via 200 principal multiplexing stations. Potentially the network can be expanded to handle more than one million users.

收起

On Distributed Communications:

IX. Security, Secrecy, and Tamper-Free Considerations

by Paul Baran

更多

One in a series of eleven Memoranda detailing the Distributed Adaptive Message Block Network, a proposed digital data communications system based on a distributed network concept. It considers the security aspects of such a system, in which secrecy is of great importance. Present security concepts are based on an implied assumption that any "cleared" person must be trusted and that any "uncleared" person is a potential spy. Further, information is either classified or not. From time to time one wonders if these binary attitudes are really a valid basis on which to predicate a military communications systems. This Memorandum, in which the underlying concepts and resulting safeguards to be built into the network are described, is written on the basis of fully anticipating the existance of spies within our ostensibly secure communications secrecy protection structure; hence, our primary interest is in raising the price of espied information to an excessive level.

收起

On Distributed Communications

VIII. The Multiplexing Station

by Paul Baran

更多

One in a series of eleven Memoranda detailing the Distributed Adaptive Message Block Network, a proposed digital data communications system based on a distributed network concept. It describes the multiplexing stations that connect subscribers to switching nodes (detailed in RM-3763-PR) of the proposed broadband distributed network. The description is presented in engineering detail, showing how the station can process simultaneously traffic from as many as 1024 separate users, sending a mixture of start-stop teletype signals and synchronous signals at various rates.

收起