The Complexity Odometer: A New Look at Preventive Maintenance and Fleet Management



Armed forces rely on fleets of vehicles, aircraft, ships, submarines and other equipment which needs to be deployed in a very wide range of conflict scenarios, under broadly varying conditions, climates, terrains, and which need to be maintained and serviced over periods spanning years, often decades. Maintenance costs are generally high and tend to increase with the complexity of the equipment.

Evidently, the frequency and cost of maintenance depends also on how intensely or severely a particular piece of equipment has been used. This impacts the accumulated wear and tear (aging) or deterioration of the equipment.

Preventive Maintenance¬†(PM) is “a routine for periodically inspecting” with the goal of “noticing small problems and fixing them before major ones develop.” There exist various forms of Preventive Maintenance. In this blog we wish to illustrate a novel approach to preventive maintenance which has a more general and, as usual, a systemic flavor.

Modern military equipment carries a multitude of sensors which monitor all critical components and subsystems, producing large amounts of real-time data. This data may be processed to measure the complexity of a particular subsystem or even the entire system. Complexity is a measure of structured information that is produced by a sensor array and provides information on the “overall level of activity” – if the data is rich and the various data channels are highly correlated the system in question is more “stressed” than when complexity is low. sudden complexity spikes very often anticipate problems, especially those that are hidden but which, when unnoticed, may lead to systemic collapse and breakdown. The technology to accomplish a real-time monitoring of complexity – which is, for all practical purposes, a new and insidious form of systemic risk – has been developed by Ontonix and SAIC and looks like this:


However, it is possible to go beyond just delivering early warnings of systemic software issues or malfunctions. Complexity may also be used to measure the accumulated “amount of activity” of a system/mobile platform or a fleet thereof. This is accomplished very simply by integrating complexity over time:


where AA stands for Accumulated Activity. While complexity is measured in bits, Accumulated Activity is measured in bit*seconds. An example of complexity and of its integral is shown below. It may be noticed that the integral grows in a linear fashion but there are intervals in which this growth is more or less pronounced.


The slope of the AA plot is proportional to the intensity with which the mobile platform is being operated and impacts its life span and the frequency with which maintenance must be made.

The AA metric – which may be called the Complexity Odometer – is similar to the kilometer count in a car. This determines when one must change engine oil or replace the tyres. Clearly, a car that is driven in a city (taxi) versus one that cruises on motorways will age differently and will require a different maintenance schedule. However, using the traveled distance in order to plan maintenance is a bit simplistic, especially when we speak of complex equipment such submarines, aircraft, tanks, helicopters or armored vehicles. Using the huge amount of data produced by the on-board software and electronic systems allows to take a more holistic look at how the equipment has been stressed. This is exactly what the AA metric accomplishes. What is lacking today is a single systemic indicator of what a particular system has been doing during its lifetime and of the intensity of its aging and deterioration. This is what the Complexity Odometer is for.

Below is an example of a small fleet of 8 mobile platforms which has been monitored over a certain period.



It is clear how 5 our of 8 platforms have a very similar level of AA, one (yellow curve) is slightly lower, while two (green and brown curves) have a decisively higher level of accumulated “stress”. Therefore, platforms 6 and 8 will probably require a more attentive examination upon return to base. Where to look is another key question that is easily answered. Complexity profiling – which may be done over any desired period of time – once a month, on a daily basis, once a year, or after each mission – provides information such as this:


A complexity profile shows which components have “participated” the most to the overall value of AA over the mentioned time horizon. Clearly, this information can augment or even drive the classical maintenance program.

Many applications of the Complexity Odometer are possible. One that comes to mind is to monitor soldiers on a battlefield based on smart clothing and body sensors. A possible civilian application is that of expensive medical equipment, such as MRI, PET or CAT scan machines that are monitored remotely.

On a final note – this is one is for engineers only! – information such as the AA history and the Complexity Profile, may be used in the design of next generation systems.





Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s