Learning And Leading: Distributed Architectures And The Sustainability Challenge

Publié le jeudi 26 novembre 2015

Billet de blogue

John F. Kennedy once remarked that learning and leadership are indispensable to one another.

That leaders must be able to learn, and learn quickly, is unsurprising.  In any undertaking, the ability to make course corrections in response to encountered challenges or opportunities, is crucial.  Limited learning capacity means limited adaptive capacity.  And as Darwin noted in his On the Origin of Species By Means of Natural Selection, if one doesn’t adapt, one doesn’t survive.

Overcoming challenges and seizing opportunities requires more than self-learning.  It also requires that one inspires learning in others, and exploits that which others learn.

Why?  Simply because the grand challenges of today are wicked problems. These are problems that have indefinite and labile formulations; for which the overarching problem stems arises from the dependencies among component problems in multiple sub-domains; for which values are contested, and for which every solution gives rise to other problems.

Indeed, many of today’s grandest challenges are superwicked problems: wicked problems for which time is running out, for which there is no central authority, and for which those seeking solutions are part of the problem.

The insights gleaned from trying to solve hard computational problems can be directly applied to wicked problems.

Decades ago, computer scientists hit upon distributed processing as an architectural solution to hard computational problems.

In distributed architectures, system learning and adaptation is accomplished through the coordinated activities of large numbers of independent processors or nodes, each one beavering away on a particular problem element.   

This coordination is accomplished by master nodes that receive and integrate information from peripheral nodes.   Synthesized information is then relayed back to peripheral nodes, which use the information packets to enhance their autonomous exploration of local neighbourhoods in solution space.

The continuous 2-way trafficking of information between master and peripheral nodes enables the network, as a collective, to develop candidate solutions and evaluate their effectiveness iteratively. 


The result?  Over time, the network explores large regions of solution space and converges to satisfactory solutions.


In this process, master nodes are critical.  They must know enough about the problem subdomains to be able to accurately and rapidly assess information from peripheral nodes.  They must synthesize incoming information, and use these syntheses to stimulate and guide exploration by peripheral nodes.    And they must exercise adaptive regulation, whereby control for certain processes may be transferred among master nodes, or even to peripheral nodes.

Environmental sustainability is the poster child of superwicked problems.  Finding and implementing solutions requires a distributed architecture, for which the master nodes are critical.

Master nodes in distributed sustainability networks must be familiar with multiple  solution neighbourhoods: science, economics, law and policy.  They must be able to integrate and synthesize information received from peripheral nodes in these domains.  And they must be able to communicate this synthesized information in a manner that inspires innovative exploration of, and learning from, these neighbourhoods.

In the Master's of Environmental Sustainability at the University of Ottawa, our goal is to provide students with the knowledge, skills, capacities and attitudes required to become master nodes in distributed sustainability networks.

Or, put another way, to lead by learning, and learn by leading.



- Scott Findlay is the Graduate Program Director at the Institute of the Environment. 

Haut de page