Abstract: Advanced navies, including the Indian Navy (IN), have adopted the Naval Operational Planning Process (N-OPP) format for operational planning. This process allows simultaneous, parallel planning at different echelons, facilitates contingency planning, and enables coordinated preparation even during plan development. Currently, most navies execute this process manually. However, leveraging Artificial Intelligence (AI) and Big Data Analysis could significantly enhance efficiency. AI can analyse critical aspects such as Environmental Scan, Factor Analysis, Critical Strengths/Weaknesses, Exposure Analysis, and Development of Enemy Course of Action (ECOA). The proposed approach aims to supplement planning staff efforts by allowing AI to handle sub-processes while maintaining human decision-making. An AI-Big Data Engine was successfully designed and tested. Additionally, the concept suggests real-time input acquisition from multiple agencies and continuous scanning, enhancing preparedness during contingencies and evolving situations.

Introduction

The art of war encompasses the strategic, operational, and tactical dimensions. The domain of operational warfare is the bridge between Strategy and Tactics (Indian Navy 2009).1 While operational art is a useful framework at all levels of warfare, the ensuing planning process is generically known as the Operational Planning Process (OPP). Milan Vego elucidates that, tactical victories, no matter how many and how decisive, are ultimately useless if fought outside some larger framework.2 The Naval Operational Planning Process (N-OPP) is a structured procedure used by naval forces to develop and execute military operations at sea. It involves various steps and considerations to ensure mission success, operational effectiveness, and the safety of personnel.

Artificial Intelligence (AI) and Big Data Analysis, have revolutionised data processing, analysis, and utilisation. AI involves the development of computer systems that perform tasks requiring human intelligence, such as understanding natural language, pattern recognition, decision-making, and learning from experience.3 AI systems can be categorised as Narrow AI, designed for specific tasks, and General AI, capable of many intellectual tasks a human can do. On the other hand, Big Data Analysis (BDA) focuses on deriving insights from large volumes of structured and unstructured data. The integration of AI and BDA has significantly advanced various industries. These technologies offer great potential for enhancing naval operational planning by providing valuable insights through advanced algorithms and data analysis, enabling more efficient decision-making by naval planners. This essay looks at how certain sub-parts of this process may be better ensured.

Analysis of Naval Operational Planning Process (N-OPP)

Planning assumes a very important role in Naval Operations. At any level, planning is derived from guidance and directives from higher levels and thus flows downstream through hierarchical echelons. Planning is undertaken by the Commander and his staff to have a detailed insight into the domain, situational awareness, responses expected from the enemy and implication of own responses. Planning hence states the Desired End State (DES), the path to be taken and routing details of the path (decisive points/ DPs). The planning culminates in the issuance of detailed operational orders. Within this framework, this study focuses on the planning undertaken at the Operational/ Headquarters level by Command Staff. The Indian Navy now uses the Naval Operational Planning Process (N-OPP) for operational-level planning.4 The US Navy, utilises a similar procedure called the Operational Planning Process.5 Components of N-OPP are enumerated in succeeding paragraphs.

N-OPP undertakes ‘Review of Situation’ and also an ‘Environment Scan.’ These sub-processes analyse the enemy’s capabilities as well as internal dynamics and is undertaken by analysing various factors like Political, Economic, Military, Diplomatic, Social, Communication/ Information, Infrastructure/ Technology, Legal/ Ethics and the Cyber Domain, better recognised by the simple and shorter acronym of DIME (Diplomatic/ Information and Intelligence/ Military and Economic). Mission Analysis is the stage where the Commander provides direction to his staff and a clear Statement of Mission is drafted. Various tasks, assumptions, constraints, and restraints are discussed and own strategic as well as operational Centre(s) of Gravity are determined. Factor Analysis is an important sub-process of N-OPP, wherein factors of Force, Space and Time are derived and examined. The interplays between Force-Time, Space- Time and overall implications are also analysed.

During the Feature & COG Analysis stage, the Critical Strengths and Weaknesses of the enemy are collated and broken down through various procedures to derive Critical Requirements, Critical Vulnerabilities and Critical Capabilities. The Opportunities, Threats, Shortfalls and Risks are analysed during the Exposure Analysis stage, for our own forces as well as the Enemy. An important sub-process is the generation of Own Courses of Action (OCOA) as well as the Enemy Courses of Action (ECOA). The process also generates CONOPS, lines of operation, phasing – sequencing – synchronisation of operational and culminates with promulgation of detailed orders.

Introduction of AI and Big Data Analytics in Operational Planning Process

The tenets of the operational planning process have been briefly described in the above paragraphs to help a transition to the main discussion. The process is manually intensive and often commences because of a trigger, either to a contingency or a developing situation. The entire process is presently being undertaken manually in most Navies. Certain aspects of the Operational Planning Process could be processed using Artificial Intelligence (AI) and Big Data Analysis (BDA), which would reduce human effort so that the operational planning teams could utilise these benefits to develop the Estimate of the Situation and the consequent concluding steps of carefully working out possible ECOAs and OCOAs.

As a part of the present study, the process has been analysed to identify the sub-processes that could be driven through AI as well as BDA Environmental Scan, Factor Analysis, Critical Strengths/ Weaknesses, Exposure Analysis and Development of Enemy Course of Action (ECOA) are some of the sections which could be analysed more expeditiously with the help of AI and BDA. The AI engine that will supplement the operational planning would be a continuous real time engine perpetually scanning the environment. AI augments the decision-making ability of humans. While the manual procedure is limited by the values/ limits set by the user, AI continuously modifies the relationship between input data and output.6 The operational planning deals with huge amounts of data and complex interaction between various factors. In such a scenario the AI and BDA can support a Commander in his decision making. Thus, the use of AI, in the sub- sections of planning could ease the planning load of staff. This would also enable the planners to focus on critical areas of the process, without wasting precious time on data collation, environment scan and other modular functions.

Case Study of a Commercial Model (Created by M/s 3M Frames7)

A case study was undertaken, on corporate processes, where AI & BDA were utilised for predictions. Stock predictions are routinely and successfully being undertaken based on extensive media scan. The architecture of AI is suitable for these functions. The media scan could be equated to Environment Scan, in military planning. The analysis also could be extrapolated to Factor Analysis, Analysis of Strengths & Weaknesses and Exposure Analysis. Stock markets are highly dynamic and unpredictable. The stock market is vastly influenced by politics, geo politics and environment events. Environment Scan based on AI could be used to predict stock market fluctuations to a large extent. This analysis is a part of Natural Language Processing (NLP) techniques, wherein, a scan of media, social media post and news clippings are analysed to gauge overall sentiment to the stock of a particular company.

This author undertook a practical analysis of an existing stock prediction live AI model in conjunction with a private firm, M/s 3 Frames Lab. The firm undertook a study for stock prediction with respect to the stocks of Apple, Yahoo and Google for an international credit firm (name withheld). The stock predictions were undertaken based on extensive media scan. This could be equated to environment scan in military planning. Data was derived from media, news clippings as well as social media scans through a customised AI engine. The data was analysed using BDA by the AI engine to identify underlying patterns. These patterns were then utilised to predict and extrapolate behaviour of stock tickers. This enabled the firm to redeploy/ re-invest stocks as per the predictions. News articles and tweets were processed, both for contents as well as sentiment analysis. A word cloud was created by the software and the cloud was mapped to word articles. News articles were also searched by words as well as stock symbols and indicators. The indices and symbols were also predicted by the AI engine. The case study established that AI & BDA had been successfully utilised for planning in commercial/ corporate sectors.

Development of AI Engine Prototype for Operational Planning

Concept Flow Diagram of AI Engine. The Case Study undertaken on stock market predictions, clearly encouraged the extrapolation of stock market planning model by AI and BDA to the N-OPP also. The Flow Diagram of AI Engine proposed for Operational Planning is placed at Fig 1.

Fig 1: AI Engine Flow Diagram (Source – Author)

The AI Engine would be designed to cater to the following as inputs:

  • Global Scan. The AI Engine undertakes a detailed media and social media scan to gauge the political scenario, economic status, social issues & fault lines, Technological developments including nuclear, cyber, space & IT domains. Other aspects like strength of judiciary, legal system, environment, energy sources including reserves, renewable sources as well as green technology are also scanned.

  • Net Assessment. The Net Assessments and Intelligence inputs are conjoined and compared to generate SWOT Analysis. The Net Assessments regularly undertaken at the College of Defence Management could provide vital input to the AI Engine.

  • Op Factors & Interaction. Op Factors and interaction involves factors of Force, Time and Space, as well as their internal interactions. The inputs to the AI Engine will be provided by the Intelligence inputs, including those from the Defence Space Agency towards analysis of terrain for space- time calculations.

  • Critical Strength & Weaknesses of Enemy. Critical strengths and weaknesses of enemy are computed by AI gleaned from inputs of net assessment, media scans and intelligence inputs. These inputs could be subsequently utilised by planners to generate COG analysis so as to derive critical capabilities, requirements and vulnerabilities.

  • Exposure Analysis. Exposure analysis highlights various threats posed to own forces, which form opportunity to the enemy. Similarly, the threats to enemy pose opportunity to one’s own forces. Both factors are analysed to compute exposure analysis (opportunities/ threats) by inputs from media scan, intelligence inputs and net assessments.

  • Enemy Courses of Action (ECOA). A large number of think tanks work independently and frequently wargame various scenarios internally. Think tanks could be synergised and networked. All inputs could be utilised as data by the BDA Engine to generate possible ECOA. Wargame outputs and lessons learnt from various organisations like Maritime Warfare Centre (MWCs), Defence Services Staff College (DSSC) and Army/ Navy/ Airforce War Colleges, could also be provided as inputs to the Big Data Analytics of AI Engine to generate possible ECOAs.

Data Inputs for the AI Engine

  • Content Providers. Contents will be collated and aggregated by the AI Engine from multiple sources as follows:

    1. Social media - Facebook, Twitter (X), Instagram feeds, etc.

    2. News Aggregators/ Syndicates, news outlets.

    3. Open-source intelligence and inputs from intelligence agencies as well Open-Source Intelligence (OSINT).

    1. Think-tank inputs could be collated by integrating outputs from various think-tanks.

    2. War Game Outputs from MWCs, DSSC and Army/ Navy/ Airforce War Colleges.

    3. Net Assessment by Military as well as civilian agencies.

    4. Inherent data bases held with Commands.

  • War Gaming Inputs. War games are regularly undertaken by all MWCs at various Commands, War Colleges (NWC, AWC and CAW), CDM and DSSC. Moreover, war games are also undertaken at MWCs as part of Theatre/ Command Level Exercises. Lessons learnt are also promulgated post major exercises. Outputs, lessons learnt, Own Courses of Action, Enemy Courses of Action etc under various scenarios could be consolidated as input to the AI Engine for Big Data Analysis.

  • Think-Tanks. Military Think tanks undertake various region-based (West Asia Dynamics, Red Sea Situation etc) as well as situation-based (Israel-Hamas War, Russia-Ukraine Conflict etc) analyses. War games are also undertaken by these think tanks. Inputs from military think tanks could be amalgamated by Service HQs/ HQIDS and be passed on as inputs to the AI Engine.

  • Net Assessment. CDM undertakes net assessments of various countries regularly. Even during the net assessment of a given country, analysis of regional, near-regional and global neighbours of that country is undertaken through a PESTLEM (Political, Economic, Social, Technology, Legal, Environment and Military) framework. Net assessment input of the main country and PESTLEM analysis of associated countries could be inputs to the AI Engine for SWOT Analysis.

  • Inputs from Intelligence Agencies. Inputs from intelligence agencies can be collated (at HQIDS or NSCS) and could be provided to NWC as inputs to AI Engine and for BIG Data Analysis. DSA inputs are also required to be provided to the AI Engine and so as to derive Force - Space computations.

Development of AI Engine Architecture

An AI Engine was subsequently designed and developed to cater to the proposed functionalities. The technical architecture of the proposed AI Engine is as placed in Fig 2. In the technical design, hardware components as well as associated embedded software are integrated into one single engine. The hardware/ software sub-components are subsequently selected and integrated as a consolidated AI Engine.

The AI Engine is a complex amalgamation of multiple sub-components that are integrated by various supervisory applications to achieve AI as well as BDA functionalities. The sub-components vary from programme interfaces to queue managers and large language models. The commercial names of the sub-components will vary from the firms designing the AI engine and are proprietary to the firm. In view of the above, technical (as against commercial) terminologies are used for the sub-components in the present design.

Fig 2: AI Engine Architecture (Courtesy – M/s 3 Frames)

Simulation Runs of AI Engine at Lab, Results & Analysis

Simulation runs of the AI Model were undertaken at the Lab (M/s 3 Frames). Environment scans concerning Political, Economic, Social, Technological, Legal, Environment and Military aspects of the adversary are depicted in graphs based on AI algorithms with certain impact probabilities to the current operation. The synthesised output could be used to derive the Threats and Opportunities of the enemy. The synthesised outputs are fed to the BDA module to derive the implications of the Op Factor analysis. This AI model scans the inputs and derives Strengths and Weaknesses as synthesised outputs of the AI. The exposure Analysis module scans the inputs and derives exposure/ risk to own forces from the enemy as synthesised outputs of AI. The AI Module synthesises various Enemy Courses of Action provided as input from various sources to derive a set of probable Enemy Courses of Action based on other inter-related planning parameters. Machine Learning (ML) could be utilised at this stage to complement BDA, so that the Engine continuously learns from the enemy actions and thinks like the enemy, whilst providing ECOA outputs. This will function like an AI based Red Cell, complementing the efforts of the planners.

Implementation of the Concept

The proposed implementation methodology is schematically depicted in Fig 3. It is proposed that the AI Engine could be installed at the Maritime

Fig 3: Proposed Methodology for Implementing the Process (Source – Author)

Doctrines and Concepts Centre (MDCC), for a continuous real-time scan and analysis in offline mode. One online module will continue to undertake a real-time Global Scan to provide synthesised data to the offline AI Engine server. A server is also recommended to be installed at Naval War College (NWC) for research by young Naval Officers undergoing courses, to provide innovative inputs to the AI developers to improve the model. The real time AI and BDA engine could, therefore, continuously deliver data for naval planners. In the event of a contingency, the concerned crisis-response planning Commands could requisition relevant data from MDCC as initial planning inputs.

Way Ahead

The concept tentatively tested by the prototype as a part of the study, now needs to be further tested on a larger scale with real-time data to populate the engine for BDA The organisation as proposed could be implemented in phases on a trial basis. The output of AI and Big Data Engines could be used in the planning of minor war games and subsequently extended for the planning of Theatre Level Exercises. The model once tested and refined could be then utilised for both crisis response and contingency planning.

Such models could, obviously be tweaked to monitor, for example, internal security flashpoints. The real-time AI and BDA could undertake continuous scans of social media, news clippings and other media outlets. The AI Engine could provide synthesised outputs and impact probability concerning possible flashpoints.

Conclusion

The operational planning process is manually intensive for likely contingencies or commences at a trigger – either a new form of contingency or a developing situation. The entire process is presently being undertaken manually in most navies. Certain aspects of the Operational Planning Process could be processed using Artificial Intelligence and Big Data Analysis, which would reduce human effort, and assist the planners to focus on other operational aspects. This study pertains to planning undertaken at Operational/ Headquarters level by Command Staff.

The concept of the study aims at supplementing the efforts of planning staff (Command Headquarters/ operational level) in undertaking various sub-processes by AI/ Big Data Analytics, thus leaving the planners with more time as well as the mind-space to use their expertise to utilise the outputs of these processes and then develop estimates of the situation as required for final human-owned Enemy and Own COAs. This will also ensure that no aspects are missed in the planning process and human supervision of the critical stages remains paramount.

Endnotes

  1. Indian Navy, Indian Maritime Doctrine (2009), 150.

  2. Milan Vego, Operational Warfare at Sea (New York: Routledge, 2017), 1.

  3. J Bindra, The Tech Whisperer (Gurgaon: Penguin Random House, 2019), 203.

  4. Indian Navy, Indian Maritime Doctrine, 150-55.

  5. US Navy, NWP 5- 01, Naval Operational Planning ( 1988), https://digital-commons.usnwc.edu, accessed 20 Sep 23.

  6. LC Davis, “Artificial Intelligence at Operational Level of Warfare,” Defense & Security Analysis, 38(1), 74–90, https://doi.org/10.1080/14751798.2022.2031692 , accessed 07 July 2024.

  7. Mr Ramakrishna, M/s 3 Frames. (2023, Sep 10). Use of AI in Commercial Fields. (C. L. Francis, Interviewer).