Digital technologies are enabling us to develop systems with huge numbers of interconnected components and sophisticated software that infuses them with seemingly unlimited capabilities. They are penetrating just about every nook and cranny of the economy and of society in general. And, they are profoundly changing the way all organizations operate, as well as our working and personal lives.
We are increasingly able to pull all these capabilities together and develop sociotechnical systems, which combine the powerful, inexpensive and ubiquitous digital technologies with the people and organizations they are transforming. But, we have much to learn on how to best design and operate these systems. They generally exhibit a level of complexity that is often beyond our ability to understand and control. Not only do they have to deal with the complexities associated with large scale hardware and software infrastructures, but with the even more complex issues involved in human and organizational behaviors.
A few weeks ago, MIT professor Joseph Sussman addressed these issues in a lecture - Complex Sociotechnical Systems: The Case for a New Field of Study. Sussman is Professor of Civil and Environmental Engineering and Engineering Systems and interim director of the MIT Engineering Systems Division.
Complex sociotechnical systems are high-impact, difficult-to-understand, technology-intensive systems that have major societal, political and economic implications. While this definition can rightfully apply to many of the systems we have been developing in the last few decades, their importance is on the rise now for two major reasons.
We are increasingly using systems thinking to address some of our most critical contemporary problems including climate change, energy and the environment, cities and metropolitan areas, jobs and productivity, the global economy, financial systems and national security. At the same time, we now have the technologies, tools, engineering methodologies and scientific principles to be able to better design and manage such systems.
In his lecture, which can be seen here, Professor Sussman says that it is not enough for us to study complex sociotechnical systems. We are not just passive observers. We are actually developing such systems in a variety of domains, so it is important that this new field of study should also be prescriptive in nature, helping us design better performing systems, as is the case with other engineering disciplines.
This new field is multidisciplinary in nature, and should enable us to address a particular domain, - e.g., transportation, energy, finance, - by bringing together a number of different methodologies. These include: systems-oriented techniques, like optimization, stochastic systems, systems simulation and systems dynamics; social sciences, management and planning, including economics, sociology, organizational behavior and strategy; and deep quantitative engineering science, which brings all the methodologies together and focuses on design, development, testing and operations.
But, asks Sussman, is there enough intellectual content in complex sociotechnical systems (CSS) to justify the creation of a new field of study? “Are there principles and core underlying concepts for creating this integrated approach across domains?” Yes, he argues in the lecture and in this working paper, where he lists some of the major CSS concepts. Let me discuss a few.
Complex systems are generally composed of different kinds of components, organized into intricate subsytems, all highly interconnected and interacting with each other. A major area of study is the relationship between the behavior of the individual components and the macro-behavior of the CSS. We need to understand how the interactions of the various subsystems affect each other, since generally, it is these interactions that account for the dynamic, unpredictable behavior of the overall CSS. Given the complexity of the systems involved, we need to understand when high level abstractions are adequate to understand the behaviors and interactions of the various subsystems, and when we have to go deeper into their various components.
In the lecture, Sussman illustrates these concepts with a discussion of the transportation issues in Mexico City, one of the world’s largest megacities. Mexico City has major transportation problems, which have serious productivity and economic consequences. At the same time, the city has severe environmental issues which affect the health of its residents. There are also social equity issues surrounding the transportation costs, which are a burden to many of the city’s residents. You cannot formulate a reasonable transportation strategy for Mexico City, or any other metropolitan area, without taking into account all these various issues.
Furthermore, the various components of a CSS often operate at very different time scales. For example, lowering the costs of public transportation addresses the social equity issues, but could result in serious overcrowding. The solution is to build additional subway and light rail lines, but that is expensive and takes a long time. A nearer term solution to overcrowding is to increase the number of buses, but that could negatively impact the environmental issues, unless the buses are electric, which then requires a new support infrastructure.
In addition, the multiple stakeholders of a complex sociotechnical systems often have disparate interests and thus quite different goals and objectives, as well as different ways of evaluating system performance. Sussman calls this evaluative complexity. For example, adding a runway to a crowded airport will reduce delays and potentially increase the number of flights to the city. While this will benefit most city residents, those living near the airport will be subject to additional noise as a result of the extra flights. Trying to satisfy the various stakeholders adds another set of difficulties to those already faced by the system designers and operators.
There is no such thing as an optimal CSS design. Instead, one must generate a set of good-enough, strategic alternatives, carefully evaluate their feasibility, and then select the one that best satisfies the various requirements and stakeholders. We must approach the design from a pragmatic point of view based on satisficing, - not optimizing.
All these various issues have to be accounted for in a reasonable CSS strategy. Sussman believes that sustainability should be the overarching design principle for complex sociotechnical systems as defined by the three Es: Economic development, concern with the Environment and social Equity.
When dealing with CSS designs, we have much to learn from biological systems. As in biology, the key attributes of a good CSS design are robustness and flexibility. The system must be robust enough to perform adequately under lots of different conditions, including failures of individual components, unanticipated interactions and external events, and a wide range of human decision making styles and choices. And, the system must be flexible enough to quickly adapt to a fast changing environment.
But, in both evolution and human designs, robustness and flexibility come at a price. The sophisticated control mechanisms needed to increase robustness and flexibility, will in turn make the system considerably more complex and add their own unanticipated failure modes. These will then be corrected over time with additional robust mechanisms, which then further add to the complexity of the system, and so on. This balancing act between complexity and robustness is never done.
For example, one of the most important protection mechanisms in biology is the immune system, which guards against disease. But the immune system is subject to its own serious diseases, such as immunodeficiencies when its activity is abnormally slow, and autoimmunities, which are caused by a hyperactive immune system.
Similarly what we often referred to as the ilities in a well designed IT system, - reliability, serviceability, availability, scalability, usability, security, disaster recovery and so on, - add significantly to the size and complexity of the system, and often leads to new kinds of design bugs.
As we continue to apply our increasingly powerful technologies to a wider range of critical problems, we will design ever more sophisticated and complex sociotechnical systems. To help us do so, we need the appropriate research and education programs.
But, in my opinion, the most powerful case for a new field of study lies in Professor Sussman’s admonition that dealing with complex sociotechnical systems requires humility on our part. We must be aware of the considerable difficulties in predicting system behavior and in making decisions that affect that behavior. And, we must keep in mind this quote from Laurence Peter, - of Peter Principle fame: “Some problems are so complex that you have to be highly intelligent and well informed just to be undecided about them.”