Emergent Behavior and SCM Introduction In this exercise, the .docx
- 1. Emergent Behavior and SCM Introduction: In this exercise, the student will analyze emergent behavior as it applies to SCM. Tasks: Read "Executive Insight in Hugos": Essentials of Supply Chain Management, answer the following questions: • Explain how negative feedback improves the performance of a supply chain. • Describe the steps that managers can take to encourage positive emergent behavior in their supply chains. • Why is emergent behavior important to continued success? 2-3 pages. APA citations. Emergent behavior is what happens when an interconnected system of relatively simple elements begins to self-organize to form a more intelligent and more adaptive higher-level system. Steven Johnson in his book, Emergence: The Connected Lives of Ants, Brains, Cities, and Software, explores the conditions that bring about this phenomenon. In an interview with Steven Johnson I posed six questions and asked him to share his insights on a range of topics. These topics range from what gives a system emergent characteristics to how could companies organize their supply chains so as to encourage and benefit from emergent behavior. · What is an “emergent system”? How is an emergent system different from an assembly line? The catchphrase that I sometimes use is that an emergent system is “smarter” than the sum of its parts. They tend to be systems made up of many interacting agents, each of which is following relatively simple rules governing its encounters with other agents. Somehow, out of all these local interactions, a higher-level, global intelligence
- 2. “emerges.” The extraordinary thing about these systems is that there's no master planner or executive branch—the overall group creates the intelligence and adaptability; it's not something passed down from the leadership. An ant colony is a great example of this: colonies manage to pull off extraordinary feats of resource management and engineering and task allocation, all by following remarkably simple rules of interaction, using a simple chemical language to communicate. There's a queen ant in the colony, but she's only called that because she's the chief reproductive engine for the colony—she doesn't have any actually command authority. The ordinary ants just do the thinking collectively, without a leader. A key difference between an emergent system and an assembly line lies in the fluidity of the emergent system: randomness is a key component of the way an ant colony will explore a given environment—take the random element out, and the colony gets much less interesting, much less capable of stumbling across new ideas. Assembly lines are all about setting fixed patterns, and eliminating randomness; emergence is all about stumbling across new patterns that work better than the old ones. · You say that such systems are “bottom up systems, not top- down.” These systems solve problems by drawing on masses of simple elements instead of relying on a single, intelligent “executive branch.” What does this mean for people who are trying to design and build emergent systems? One of the central lessons, I think, is that emergent systems are always slightly out of control. Their unpredictability is part of their charm, and their power, but it can be threatening to engineers and planners who have been trained to eliminate unpredictability at every turn. Some of the systems that I've looked at combine emergent properties and evolutionary ones: the emergent system generates lots of new configurations and ideas, and then there's a kind of natural selection that weeds out the bad ideas and encourages the good ones. That's largely what a designer of emergent systems should think about doing: it's closer to growing a garden than it is building a factory.
- 3. · What does it mean when you say that emergent systems display complex adaptive behavior? The complexity refers to the number of interacting parts, like the thousands of ants in a colony, or the pedestrians on a street in a busy city. Adaptive behavior is what happens when all those component parts create useful higher-level structures or patterns of behavior with their group interactions, when they create something—usually unwittingly—that benefits the members of the group. When an ant colony determines the shortest route to a new source of food and quickly assembles a line of ants to transport the food back to the nest; when thousands of urbanites create a neighborhood with a distinct personality that helps organize and give shape to an otherwise overwhelming city—these are examples of adaptive behavior. · What is negative feedback as opposed to positive feedback? What role does negative feedback play in the ability of a system to exhibit adaptive behavior? Negative feedback is crucial, and it's not at all negative in a value-judgment sense. Positive feedback is what we generally mean when we talk about feedback, as in the guitar effect that we first started to hear as music in the 60s: music is played through a speaker, which is picked up by a microphone, which then broadcasts it out though the speaker, creating a sound that the microphone picks up, and so on until you get a howling noise that sounds nothing like the original music. So positive feedback is a kind of self- perpetuating, additive effect: plug output A into input B which is plugged into input A. Negative feedback is what you use when you need to dampen down a chain like this, when there's a danger of a kind of runaway effect, or when you're trying to home in on a specific target. Think of a thermostat trying to reach a preset temperature: it samples the air, and if the air's too cold, it turns the heat on, then samples it again. Without negative feedback, the room would just keep getting hotter, but the thermostat has been designed to turn the heat off when the air reaches the target temperature. Ants use a comparable technique to achieve the right balance of task allocation
- 4. throughout the colony: an individual ant who happens to be on foraging duty will sample the number of ants also on foraging duty that she stumbles across over the course of an hour—if she encounters a certain number, she'll switch over to another task (nest building, say) in order to keep the colony from becoming overrun with foragers. · In your book you mention a designer who has proposed building a learning network of traffic lights that will find an optimal solution to continually changing traffic conditions. You observe that, “You can conquer gridlock by making the grid itself smart.” What is it that would make the grid smart? Is this grid an example of an emergent system? The idea proposed in the traffic model is not to take the traditional engineering, top- down approach and say: “let's look at the entire city and figure out where all the problems are, and try to design the roads and the light system to eliminate the problems.” The smart grid approach is to give each light a local perspective with a little bit of information, and give it the goal of minimizing delays at its own little corner. So the light would be able to register the number of cars stacked up at the intersection, and it would be able to experiment with different rhythms of red and green, with some feedback from its near neighbors. When it stumbles across a pattern that reduces delays, it sticks to that pattern; if the delays start piling up again, it starts experimenting again. The problem with this sort of approach is that on Day One it's a terrible, terrible system, because it doesn't yet know anything about traffic flows. (You'd have to teach it quite a bit before you could actually implement it.) But it would learn very quickly, and most importantly, it would be capable of responding to changing conditions, in a way that the traditionally engineered approach would not. That's a hallmark of adaptability. · Consider a system composed of many different companies whose goal is to provide a market with the highest levels of responsiveness at the lowest cost to themselves. High levels of responsiveness require that these companies work together to
- 5. design, make, and deliver the right products at the right price at the right time in the right amounts. What are some of the things that these companies could do to organize themselves into an emergent system? There's a telltale term in supple chain systems, which may well be unavoidable—the term “chain” itself. Almost all emergent systems are networks or grids; they tend to be flatter and more horizontal, with interaction possible between all the various agents. The problem that supply chains have with positive feedback revolves around the distance between the consumer and those suppliers further down the chain—because the information has to pass through so many intermediaries, you get distortion in the message. Most emergent systems that I've looked at have a great diversity of potential routes that information can follow; the more chain-like they become, they less adaptive they are. The other key here is experimentation: letting the system evolve new patterns of interaction on its own, since these can often be more useful and efficient than the pre-planned ones. Of course, you don't want to waste a few economic quarters experimenting with different supply chains, most of which are a disaster. But that's where some of the wonderful new modeling systems for complex behavior can be very handy: you can do the experimenting on the computer, and then pick the best solutions to implement in real life. Reference Hugos, Michael H.. Essentials of Supply Chain Management, 3rd Edition. 3. VitalSource Bookshelf. John Wiley & Sons (P&T), 2011-07-11, Sunday, August 12, 2012. <http://digitalbookshelf.argosy.edu/books/9781118279229/10/6 >