Fuzzy Logic in Support of Decision Analysis in Intelligence Warfare

22 November 2006

Fuzzy logic is an often-criticized mathematical theory, which posits that there is middle ground in many decisions and accounts for vague notions such as “very,” “a little” and “most.” Modern computer languages and decision analysis programs rely on classic mathematics and statistics that allow only for precise “true/false” statements. This “true/false” concept is evidenced in binary programming and the never-ending data streams of 1’s and 0’s found in today’s digital world. However, in cases where a statement may be both “true” and “false” at the same time, fuzzy logic can be utilized to solve the problem. Fuzzy logic uses natural language based programming such as “Bill is very tall” and “The weather is warm” to allow more human-like interpretation and reasoning where conventional mathematics would have a difficult time. Fuzzy logic is capable of using simple equations to solve complex problems, including decision processing and analysis. It provides a means for dealing with predicaments where no exact information is available and its vagueness is utilized to achieve simplicity, stoutness and a low-cost solution. In decision analysis, strategists often employ decision trees that can strictly define the inputs and assign a probability to outcomes. In support of intelligence warfare, the inputs are usually blurred and little hard data is available or it is diversified over a large geographical area, making it difficult to allocate accurately the outcome’s likelihood. Because of the nature of intelligence predictions and decision analysis, fuzzy logic can be utilized in nearly every step of the intelligence cycle and create options for planners and decision makers. Additionally, offensive intelligence warfare can benefit from fuzzy logic by efficiently applying the least amount of force necessary to achieve the desired results. For these reasons, intelligence strategists can use fuzzy logic systems to aid in management and decision analysis.

First invented as a representation method and calculus for unsure or indistinct concepts, fuzzy logic is fundamentally a multi-valued logic that permits more human-like analysis and calculations in computers by creating subcategories between concepts such as “true/false”, “hot/cold” etc. The idea for fuzzy logic is steeped in history, though Lofi Zadeh, U.C. Berkley professor, is accredited with its discovery and development in 1965.[1] Over a thousand years before Zadeh, however, philosophers such as Plato had posited the Law of Excluded Middle, yet later proposed a third area where notions of “true” and “false” coexist.[2] Parminedes propositioned the earliest version of this rule around 400 B.C. and affirmed amidst controversy that statements could be concurrently true and false.[3] This middle ground or grey area is where most of human thought and decision resides. While idealists deal with absolutes, most of the decisions humans make are arrived at through compromise, considering simultaneously multiple inputs and relying on experience and perception.

A useful allegory for considering the middle ground comes from the ancient philosophical problem of Theseus' ship. Greek legend holds that Theseus slew the Minotaur and returned to Greece with the youth of Athens. On their return, the Greeks decided to preserve Theseus' ship in perpetuity. The ship was moored and vigilantly maintained. Over the years, rotting ropes, timbers, and sails were gradually replaced with new ones. Plutarch records it thus:

"The ship wherein Theseus and the youth of Athens returned had thirty oars, and was preserved by the Athenians down even to the time of Demetrius Phalereus [~ 350–280 B.C.], for they took away the old planks as they decayed, putting in new and stronger timber in their place, insomuch that this ship became a standing example among the philosophers, for the logical question of things that grow; one side holding that the ship remained the same, and the other contending that it was not the same." [4]

The inquiry has now been debated for more than 2000 years: was the ship still Theseus' ship? Fuzzy logic theorists would answer “yes and no,” and be perfectly content with that concept. Most mathematicians would shudder at the thought of a conflicting answer; in conventional logic, the answer must be true or false (usually represented as 1 and 0, respectively), but not both, severely limiting the application of conventional mathematical equations.

Fuzzy logic academicians, conversely, assign a value between 0 and 1 (such as .5 or .23) to a statement, allowing the statement to occupy a region between true and false. This variance allows for greater application in many problems that would be difficult or impossible to solve using traditional methods. Fuzzy logic has been successfully applied in areas of air-conditioning, traffic control, handwriting recognition, automatic television picture control, and, most famously, in controlling the subway in Sendai, Japan. Dr. Zadeh is currently working on a project to integrate the Internet with fuzzy control systems. The Applications appear to be limitless and challenge conventional thought and modern computer programming, although it is the primary system employed in artificial intelligence.

Because intelligence warfare decision making involves incorporating information from a wide spectrum of sources, fuzzy logic is an ideal tool in providing information on demand, offering suggestions, and providing alerts. Dr. Zadeh contends that measured decision systems, based on conventional logic, are incapable of providing the level of support needed to make automated complex decisions. He states, “Like driving a car in heavy city traffic, humans accomplish this with no problem, however, automated driving, in heavy traffic, based on measurement, is not possible.”[5] His system takes estimates and continuously changing variables into account, relying more on experience, rather than millisecond measurements. Likewise, application of fuzzy logic into decision analysis can allow for a wider range of inputs and outcomes, weighing in abstract ideas and vague notions of danger and vulnerability.

Fuzzy logic sets can be applied to the entire intelligence cycle, beginning with planning and direction, collection, processing, analysis and production, and finally, dissemination. The planning and direction process involves the administration of the entire intelligence effort, from the recognition of the requirement for data to the final delivery of an intelligence product to the consumer.[6] Because, in some cases, requests and requirements for information have become institutionalized, a baseline of data has been acquired and a fuzzy logic system could easily identify and alert intelligence officers to changes within a certain subset, such as nuclear forces in China.

Once the request for information has been levied, the acquisition and processing of data and intelligence is often overwhelming to the technician, who must sift like a gold miner for the important data. Fuzzy applications that “learn” from the user and “adapt” to individual working styles can help to identify potential nuggets to the collector/processor. Additionally, language and voice recognition are among tasks that fuzzy logic has been shown to excel in and this function can assist in voice intercept collection. Fuzzy logic systems in networking can also balance the workload among collectors, regulating how much data “sits” in the inbox.

During the analysis and production stage, raw intelligence is converted into finished intelligence. Because the preparation of intelligence includes data that is regularly fragmentary and often contradictory, conventional computer systems have been inept at automating the tasks, leaving the bulk of the work for the human analyst. Fuzzy logic systems are an ideal match for this often vague and seemingly contradictory environment. Aided by unconventional fuzzy logic programs, analysts can use fuzzy logic systems to more easily see the relationships between disparate data and become more efficient.[7] Finally, dissemination of the finished product could become better automated, in addition to providing the report to the originator of the product, network fuzzy systems could monitor the type of reports other analysts and consumers read, and “suggest” that they read other relative reports from various sources previously untapped.

When considering the application of force against a foe, fuzzy logic can aid the decision makers by providing a non-linear forum for exploring actions and their outcomes. In conventional decision analysis, the user utilizes a decision tree, with forks and branches, presenting the user with defined options and structured outcomes, usually with a degree of probability attached to each future.[8] This is a good, time-tested model for predicting and measuring success in a known environment. Unfortunately, international relations are rarely stable and are dependant on a multitude of variables. Similarly, intelligence warfare is often described as an ongoing process of offensive, exploitation, and defensive information functions, with “overlapping … degrees of intensity moving from daily unstructured attacks to focused net warfare of increasing intensity until militaries engage in C2W.”[9] These “degrees of intensity” are not easily measured by conventional means; therefore, fuzzy logic appears to be a more appropriate model in this application. Further, the appropriate equilibrium of force determined by fuzzy logic systems can be easily controlled and automated to minimize collateral damage and maximize the benefit of intelligence and conventional operations.

While decision theory and analysis have long enjoyed the support of conventional logic systems, fuzzy logic, in support of intelligence operations, can provide the additional tools required for inputting and manipulating data to achieve a better product. By aligning computer logic systems closer to the natural human thought process, faster and more accurate results can be achieved without requiring advanced technical training from the operator. Allowing reasonable and scaled inputs into a computer model such as: definitely yes, probably yes, maybe, probably no, and definitely no, will help the analyst better describe the intelligence situation. Better descriptions of intelligence situations (though conventional logic would call them vague descriptions), aided by fuzzy logic systems to predict and adapt more quickly to non-uniform operations and occurrences, serve the decision and analysis process throughout the entire intelligence process and decision analysis.

Bibliography


Agarwal, Pragya Lotfi Zadeh: Fuzzy logic-Incoporating Real-World Vagueness, Center for Spatially Integrated Social Science, University of California, Santa Barbara. 2001

Clough, A. H., ed. Plutarch: lives of noble Grecians and Romans. Modern Library Series. Translated by John Dryden. Modern Library, New York, New York, USA. 1992.

Kantrowitz, Mark and Erik Horstkotte, and Cliff Joslyn, Answers to Questions about Fuzzy Logic and Fuzzy Expert Systems, Accessed November 28, 2006 online see: http://www.faqs.org/faqs/fuzzy-logic/part1/

Lapin, Lawrence L. and William D. Whisler. Quantitative Decision Making with Spreadsheet Applications, Thomson Learning, Inc., 2002.

Richelson, Jeffrey T. The U.S. Intelligence Community, Westview Press, 1999.


Vernon, James. Fuzzy Logic Systems, (Control Systems Principles), PDF accessed online November 28, 2006, http://www.control-systems-principles.co.uk/whitepapers/fuzzy-logic-systems.pdf

Waltz, Edward. Information Warfare Principles and Operations, Artech House, Inc. 1998.

Zadeh, Lofti A. and Masoud Nikravesh. Perception Based Decision Systems [Power Point Presentation] (UC Berkley) 22 July 2002.

[1] James Vernon, Fuzzy Logic Systems, (Control Systems Principles), PDF accessed online November 28, 2006, http://www.control-systems-principles.co.uk/whitepapers/fuzzy-logic-systems.pdf
[2] Pragya Agarwal, Lotfi Zadeh: Fuzzy logic-Incoporating Real-World Vagueness, (Center for Spatially Integrated Social Science, University of California, Santa Barbara), 2001, accessed November 22, 2006 online: http://www.csiss.org/classics/content/68
[3] Ibid.
[4] Clough, A. H., editor. Plutarch: lives of noble Grecians and Romans. (Modern Library Series),
Translated by John Dryden. Modern Library, New York, New York, USA. 1992.

[5] Lofti A. Zadeh and Masoud Nikravesh, Perception Based Decision Systems [Power Point Presentation] (UC Berkley) 22 July 2002.
[6] Jeffrey T. Richelson, The U.S. Intelligence Community, (Westview Press, 1999), 4.
[7] Mark Kantrowitz, Erik Horstkotte, and Cliff Joslyn, Answers to Questions about Fuzzy Logic and Fuzzy Expert Systems, Accessed November 28, 2006 online see: http://www.faqs.org/faqs/fuzzy-logic/part1/
[8] Lawrence L. Lapin and William D. Whisler, Quantitative Decision Making with Spreadsheet Applications, (Thomson Learning, Inc.), 2002, 52.
[9] Edward Waltz, Information Warfare Principles and Operations, (Artech House, Inc.) 1998, 17.