The knowledge gap
© Donald V. Steward 2005
Many decisions are made without realizing what knowledge is needed or may already exist within the enterprise that should have been considered. Knowledge and Knowledge Management are necessary but not sufficient to solve problems and make decisions. Retrieving information begins with a question; the question arises from trying to solve a problem. Knowledge needs to be tied to a means of relating it to the solution of problems. We will discuss such a means.
In my early days as an engineer, I came across the following problem. In designing a nuclear power plant, the designers of the heat exchanger chose the material that is typically used for such a part. While one of these engineers and a metallurgist from the same company were at a conference on the other side of the country, they happened to meet and had an informal conversation over drinks. When they discussed this decision, the metallurgist pointed out that the water in that radioactive environment would corrode that material in no time at all. But indeed that material had been chosen and the heat exchanger was already being built. The heat exchanger had to be rebuilt at the cost of many millions of dollars.
I came across this problem and many others like it when I was an engineer. But I have since learned that this same sort of problem occurs frequently in other aspects of business as well.
Knowledge Management has received a lot of attention recently. Knowledge is the repository of learning. It can provide vital competitive advantages. But it can be expensive to collect, index and retrieve. And it’s hard to motivate people to record what they have learned so that others can easily benefit from it, particularly if they don’t anticipate how others might use it. And even if the knowledge is there, it’s doubtful that others will realize it is available and retrieve it when they should to make important decisions.
The purpose of having knowledge is that it may be used to solve problems. Sometimes this has been overlooked and knowledge just collected for its own sake.
So let’s go back a step to see the role of problem solving. We ask: What is the role of a business? Let’s consider that the business of business is solving problems. First it solves its own problem of receiving revenues by providing products and services that others can use to solve their problems. Thus, the competitive advantage to having knowledge occurs when it can be used to solve problems.
If knowledge is just facts, then it by itself does not solve problems. You also need a model that puts that knowledge together in such a way that problems can be solved. One of the most pervasive of these models is cause-and-effect where we say ‘A is caused by D and E together, or by F’. Many complex systems can be described by a collaboration of people who put together a number of such cause-and-effect statements. We show a simplistic example that might be used for auto diagnosis at the end of this article.
How do we find explanations to solve problems using cause-and-effect? We typically start with a hypothesis as to what may have caused what we are trying to explain, then we use the cause-and-effect relations to deduce consequences of that hypothesis. If the consequences correspond to what we wish to explain, then we have found an explanation. But generally we have to try hypothesis after hypothesis until we find one that’s suitable. And once we have found an explanation, we don’t know how many other explanations there may be that have been overlooked.
Thankfully, it has recently become possible to write computer software that does this backwards. The computer can begin with something to be explained and find all the possible explanations within the scope of the cause-and-effect statements. Now it remains to consider what we can do with this new capability.
The first thought is to use this to develop diagnostic systems. We show such an example of a simple system to diagnose the cause of automotive problems below. But there are already expert systems that are used by repair garages to diagnose car problems. Is there any advantage to using this cause-and-effect approach instead of expert systems?
Yes there is. Expert systems require that you already have an expert that you can brain dump to write a program that will hopefully mimic the thinking of the expert. You also need someone to analyze the system to find and remove any inconsistencies. The expertise captured in an expert system is the result of the experience of the expert, which is the result of reasoning that remains in the expert’s head and might not be captured by the expert system. That makes it hard to change the system as new knowledge becomes available.
But let’s say you don’t have an expert. Then by the method we discuss here you can create an ‘expert’ through the collaboration of people who understand various aspects of the system and contribute their own cause-and-effect statements to build the expertise. And later, as new knowledge becomes available, it can be brought into the model. And the model finds inconsistencies automatically.
There are other examples of diagnostic systems that can be developed by a collaboration of people contributing these cause-and-effect statements. For example, you could create a medical diagnostic system by putting together statements that describe the human body. Not every physician can keep up with all the new literature as it occurs. But medical people who do keep up with various parts of the literature can maintain such a cause-and-effect system. Then a physician’s diagnostic system can be updated by the Internet perhaps once a week or as often as new literature occurs.
Such a system can be used to describe how a production line works and diagnose problems occurring in the production process by asking it to explain problems that are observed. Or it could be used to help find the cause of accidents, or crimes or what-have-you. It challenges the imagination to think of all the ways this could be used. Many of these applications still need to be explored.
We can use this technique to find possibilities, which include both explanations and opportunities. It is capable of coming up with new possibilities that might not otherwise have been imagined, which is what is usually meant by creativity. Not only is it useful to find ways to make something occur, but also find ways to prevent something from occurring.
The same language used to express cause-and-effect systems can also be used to find opportunities. You can say ‘I can do this if I could do both that and that, or if I could this other thing. Using a system of expressions like this, it is possible to define your capabilities and find the various opportunities available using your existing capabilities. If an existing product or service has become unprofitable, this can be used to help find how to use your capabilities to produce another product or service that might be more profitable. This could reduce the amount of downsizing.
This technique could be used in a number of imaginative ways. It could be used as a Wiki Problem Solver. Through the web, people could collaborate in building cause-and-effect models to explain various systems. Others could then use these models through the web (Active Service Provider) to discover ways for solving their own problems. The goal of science is to find explanations for what is observed. A Wiki Problem Solver could be a useful tool for scientific collaboration. Someone can propose a hypothesis that explains certain observations. Then this hypothesis can be tested to see if it can also explain additional observations.
This example has been made naively simple by eliminating all systems other than the ‘starter’ and ‘spark to the plugs’.
Figure 1 shows how the cause and effect statements are entered in the form of an outline.
Figure 2 shows the explanation given for ‘Starter Motor Runs’ but ‘NOT Car Starts’.
Figure 3 shows the cause-and-effect diagram. The solid arrows show causes that must be taken together, i.e. AND. The dotted arrows show the causes that are alternatives, i.e. OR.
Note that if the ‘Starter Runs’ then there must be ‘Electrical Power’, which eliminates the lack of ‘Electrical Power’ as a possible cause for ‘NOT Spark Gets to Plugs’. The full solution will show both those things that are wrong and those that are right. But for diagnostic purposes, the things that are right are excluded.
© 1995-2015 Problematics
Contact us via email at: email@example.com