{"id":1082,"date":"2018-11-23T12:27:16","date_gmt":"2018-11-23T17:27:16","guid":{"rendered":"https:\/\/www.martin-parrot.com\/?p=1082"},"modified":"2018-11-23T16:31:44","modified_gmt":"2018-11-23T21:31:44","slug":"function-tree-models-a-tool-for-optimization","status":"publish","type":"post","link":"https:\/\/martin-parrot.com\/en\/function-tree-models-a-tool-for-optimization\/","title":{"rendered":"Function Tree Models \u2013 A Tool for Optimization"},"content":{"rendered":"

This article was originally published on Value World, at\u00a0http:\/\/www.value-world.org\/wordpress\/?p=472<\/a><\/em><\/p>\n

The purpose of a function tree is to illustrate all the functions that a product, a process or a project[1] must do and the links between them, in order to challenge these functions and develop a better response to the client\u2019s needs. The functions are linked together in a logical way and the model resulting from this diagram illustrates what will be done by the product, starting with a statement on the mission of the product, and presenting all the functions.<\/p>\n

The best analogy we can use to illustrate a function model is to compare it to a puzzle: when you open the box of a puzzle, you see the 1000 pieces that you will have to connect to see the desired picture. It is the same thing for a function model: you start by finding the functions of the product (the pieces of the puzzle) then you assemble them to obtain the model of the product (the picture). If you are missing functions, you can add them as you build the model (like adding pieces of the puzzle that you found outside the box) and if you have functions that don\u2019t fit, you remove them (pieces of another puzzle).<\/p>\n

The function diagram does not represent the chronology of what the product does. Indeed, no chronology whatsoever is considered here, only logic. This statement is remarkably true when a diagram represents a process, where the common mistake is to place the functions in a chronological order, instead of asking what the \u201chow-why\u201d logic is.<\/p>\n

The function diagram is also not a work breakdown structure, because it does not represent any technical solution. It only shows the functions that a product will do to satisfy specific needs.<\/p>\n

Let\u2019s explain here how to build a function diagram and the many uses of it.<\/p>\n

How to Build a Function Diagram<\/strong><\/p>\n

As for any FAST diagram, the functions are linked together with a HOW-WHY logic. The HOW goes from left to right and the WHY comes back, from right to left. There is also a more subtle relation: the WHEN. It does not represent WHEN meaning \u201cat what moment\u201d, but more like \u201cwhen we realize this function, what is the underlying goal related to it?\u201d The answer to this question is located underneath the function, but a little creativity is possible here: if no place is available under the function, it can also be placed above.<\/p>\n

The other logical relation that can be represented in a function tree, as for FAST diagrams, is the AND and the OR relation.<\/p>\n

Again, as for any FAST diagram, the level of abstraction lowers from left to right, with the high level functions on the left and the low level on the right of the diagram.<\/p>\n

Finally, we can distinguish a few types of functions:<\/p>\n

Basic function: it is the essence of the product under study. It is the main reason why your client wants your product. If this function is not realized, the product is useless.<\/p>\n

Supporting function: these functions support the basic function and help sell the product. They facilitate the realization of the basic function by making it easier to accomplish, or safer, or faster \u2026 and this aids in selling the product.<\/p>\n

Constraint function: those are imposed upon the designer by others and must be respected. These functions can be brought by environmental considerations, by the market, by regulatory agencies \u2026 but they must be carried out, or else you will not be able to sell the product.<\/p>\n

With these few rules dictating how to build a function tree diagram, we have all we need to start one.<\/p>\n

Building a Function Tree<\/strong><\/p>\n

As mentioned earlier, it is easier to first identify the functions of the product that will be placed in the diagram. A few methods exist and the choice belongs to the facilitator.<\/p>\n

Many texts are available to understand how and when to use any of these methods, in order to elicit the functions of a product. These methods will identify functions that are either of a high level of abstraction that will be placed on the left of the diagram or of a lower level of abstraction, on the right of the diagram.<\/p>\n

But the first thing to place in the diagram is the fundamental mission of the product (its task). This describes the raison d\u2019\u00eatre of the product. It is often composed of more than two words, unlike functions which are described with an active verb and a measurable noun.<\/p>\n

Once the mission statement is chosen, with the client, it is placed on the leftmost side of the diagram and will be the starting point of the tree graph. It is possible to place another function on the left side of it, to better explain the context for example, but this other function will not be part of the study.<\/p>\n

\"\"<\/p>\n

On the right of the mission, we place the functions that answer the question HOW. We find the functions we want to place here either in the list of elicited functions or by creating new ones. Obviously, the first few times, it is easier to elicit functions before starting to build the diagram, but this construction process allows you to sometimes elicit more functions.<\/p>\n

The diagram is built by placing three or four functions at a time, going to the right (answering the question HOW) then by validating the links, reading from right to left, with the question WHY. The role of the facilitator is important: he or she will read the diagram as it is being built by saying: \u201cHow do we realize this function? It\u2019s by doing this function.\u201d The brain of the attendees is wired in such a way that a false reading will automatically make it react and thus, the function will be moved to a place where it is more logical and appropriate. Many computer tools exist to help building such diagrams.<\/p>\n

In a function tree, the top of the diagram is where we tend to place the basic function(s). Directly below will be the support and constraint functions, which may be placed according to the client\u2019s order of importance.<\/p>\n

Different Types of Trees<\/strong><\/p>\n

The differences between a function tree and a FAST diagram are that the function tree will not be focused on the \u201ccritical path\u201d of functions, and also that there will be no functions \u201cfloating around\u201d, as the one-time function and all-the-time functions do. Everything is attached.<\/p>\n

As for the difference between the function tree and the customer function model (client-oriented diagram), the four functions the client diagram uses (satisfy user, attract user, assure convenience, assure reliability) may be replaced by whatever function is more suitable. Indeed, these four functions may be appropriate when dealing with a product, but other functions may be more relevant if we are describing here a service or a process or another product which does not deal with attractiveness, for example. The function tree model is very flexible.<\/p>\n

The following example will illustrate the whole process.<\/p>\n

Example of a Function Tree<\/strong><\/p>\n

This diagram represents the function model of the drum system of a clothes dryer.<\/p>\n

Let\u2019s read it together:<\/p>\n

We can see that the mission of this system is to \u201cdry content in an efficient, pleasant and safe way\u201d (in the leftmost box). But when this mission is realized, the drying process must be done by \u201cresisting operating conditions\u201d (which may vary) and by \u201crespecting standards\u201d. These two functions, when placed upstream of a tree, are true for all the functions that follow. Placing them there avoids repetition at the subsequent levels.<\/p>\n

\"\"<\/p>\n

By moving to the right, we answer the question, \u201cHow can we dry content, in a safe, pleasant and efficient way?\u201d The answer is: by drying the content AND by ensuring safety AND by ensuring reliability AND by satisfying user AND by facilitating maintenance. We can see here that function\u00a01, \u201cdry content\u201d, is the basic function of the product. Indeed, if a dryer does not dry the content we place inside, it doesn\u2019t do its job and it is thus useless. The dryer is designed and therefore bought essentially to accomplish this function, but it must also do it safely, reliably, pleasantly and it must be able to be serviced. Those are all important selling functions. Thus, these functions\u00a02 to 5 are supporting functions: they support the basic function by making sure the dryer not only dries the content but does it safely, reliably\u2026 More precisely, function\u00a02, \u201censure safety\u201d, would be a constraint function: it is vital that the product be safe and, as the designer and the supplier of this product, all must be done to achieve this goal. Functions\u00a03, \u201censure reliability\u201d, 4, \u201csatisfy user\u201d and 5, \u201cfacilitate maintenance\u201d, are support functions, because when a product is reliable, it is a great product to use over and over; satisfaction comes from a quiet use and a pretty product and finally, when the product is easy to repair, the client is happy to have it back in use quickly. All those could increase the sales.<\/p>\n

In function 3 \u201censure reliability\u201d there is a function called \u201cresist efforts\u201d. Indeed, all kinds of efforts are applied to the drum system: weight, mechanical stress, torsion, temperature, even humidity (corrosion). The system must be designed to resist all those which are defined in the functional performance specification. For example: weight of clothing, x pounds, temperature: minus 40 to plus 40 (in the shipping truck), use temperature: y degrees and so on. By defining all the efforts the system must resist, the designers can dimension the components appropriately and select the right materials, in order to have a reliable system.<\/p>\n

Let\u2019s extract a branch of this function tree model to have a closer look at it. Each branch is called a \u201cfunctional group\u201d and it contains functions that belong together.<\/p>\n

\"\"<\/p>\n

This branch of the tree illustrates in detail the basic function of \u201cdrying content\u201d. How<\/em> do we dry the content? The answer is on the right: the product must \u201ccirculate air in the content\u201d (thus removing moisture), it must \u201cconfine the content\u201d (it the content escapes, it will not be dried), and \u201cstir the content\u201d (if it stays in a heap on the bottom, it will be difficult to dry). Thus, the three functions must be realized to \u201cdry the content\u201d of the dryer. If we circulate air but we do not confine the content, it will spill out and won\u2019t dry. If we confine it but we do not stir it, the clothes on top will dry but not those underneath. Therefore, to dry the content, the product must be able to concurrently circulate air AND confine content AND stir content. If one of the three functions is not realized then the content will not be dried as expected.<\/p>\n

If we continue reading to the right by taking branch\u00a01.3, \u201cstir content\u201d, we can see: How<\/em> do we stir content? The answer is on the right: by \u201callowing rotation\u201d and by \u201ctransferring the movement\u201d to the content. Here, we have a WHEN function which can be read as follows: when<\/em> we allow rotation, the product must be able to resist the mass of the content tumbling down. Indeed, when the container turns and applies this movement to the content, the system must resist the full weight of the mass of wet clothing falling to the bottom of the container.<\/p>\n

If we read from right to left, we have: why<\/em> do we \u201callow rotation\u201d of content and \u201ctransfer the movement\u201d? To stir it. Why<\/em> de we stir it AND confine it AND circulate air? To dry the content.<\/p>\n

And so on. Note some of the functions in this actual example use more than two words. Also some are actually \u201cactivities.\u201d This example responded to a client request to include the information as shown.<\/p>\n

When is a Function Tree Finished?<\/strong><\/p>\n

Once all the functions previously elicited and those added on the way are placed, the diagram may be considered finished. More precisely, when each branch ends with a function for which no more answer can be given to the question HOW (going to the right), other than naming the solution to perform the function (a component, a sub-system, an activity, etc.), then the diagram is finished.<\/p>\n

Uses of a Function Tree<\/strong><\/p>\n

A function tree can be used in many ways:<\/p>\n