Proceedings of DETC’97:
1997 ASME Design Engineering Technical Conference
September 1417, 1997, Sacramento, California
DETC97/DTM3865
CREATIVE DESIGN METHODOLOGY AND THE SIT METHOD
Roni Horowitz
Department of Industrial Engineering
TelAviv University
TelAviv 69978, Israel
Email: horowitz@eng.tau.ac.il

Oded Maimon
Department of Industrial Engineering
TelAviv University
TelAviv 69978, Israel
Email: maimon@eng.tau.ac.il

Abstract
The paper presents SIT (Structured Inventive Thinking)  a structured method for enhancing creative problem solving in engineering design. The method is a three step procedure: problem reformulation, general search strategy selection, and an application of idea provoking techniques. The most innovative part of the method is the problem reformulation stage. The given problem is modified through the application of objectively defined and empirically tested set of sufficient conditions for creative solutions. The paper describes the sufficient conditions and the empirical study that demonstrates their appropriateness. Then the whole SIT mechanism is presented with illustrative examples.
KEYWORDS: cognitive design theories, innovative design
introduction
Inventing, as Edison said, is one percent inspiration, and ninety nine percent perspiration. Most of the existing design methods deal with the perspiration part of inventing, what is known as routine design: knowledge organization, parametric optimization, database search, team conferencing, case based design and other elements of the design process. The problem of how to support an engineer in the production of a truly creative idea is addressed by very few researchers, although the field is gaining popularity.
There are many ways to define a creative idea. Dagupta (94) counted about 80 different definitions in the literature. In this research we use a simple common definition: a creative idea is one that is considered creative by field experts. This definition is based on the assumption that people can identify a creative idea when they see one, but are unable to supply an apriori list of properties which constitute a creative idea. Using expert judges to evaluate the creativity of an idea is a common method among researchers of creativity in cognitive science (Hennessey, Amabile, 88). We use this method for the experimental evaluation of the SIT method developed here.
SIT (for Structured Inventive Thinking) is based on a theory of sufficient conditions for creative ideas (Maimon, Horowitz, 96). The theory states that if an idea for a solution of a technological problem satisfies two simple and objectively formulated conditions, that idea will be deemed creative by field experts. Using SIT, the problem solver first reformulates his problem by changing the goal from ‘find a solution’ to ‘find a solution that satisfies the conditions’. The problem solver then proceeds to the process of searching the solution. At this stage he/she selects one of two general solution strategies each leading him to a different set of two idea provoking techniques.
The study of creativity in general is different from the one developed here, and does not address design in particular. Common known general creativity strategies are Brain Storming (Osborn, 59), Morphological Analysis (Allen, 62), Lateral Thinking (de Bono, 92), and Synectics (Gordon, 61). The main differences are in the structured methodology we use and the results we get.
The development of SIT was inspired from TRIZ (Altshuller 85; Fey, Rivin, Vertkin, 94; Sushkov, Mars, 95; Malmqvist, Axelsson, Johansson, 96) which is one of the few methods for enhancing creative problem solving in engineering. The authors were introduced to the TRIZ method about 10 years ago, and since then conducted more that 100 courses and seminars in the Israel, US, and Singapore. Initially the TRIZ method was taught. Gradually, through numerous modification cycles based on experience, and through theoretical and experimental study, the method has changed to become a different method we now call SIT.
TRIZ was developed in the former Soviet Union by Genrich Altshuller. Altshuller examined thousands of inventions and patents from which he extracted some properties that characterize creative solutions. His main finding was that creative solutions incorporate an elimination of a conflict in the problem state. A conflict is a state where one parameter must be changed, in order to get some benefit, but changing that parameter causes a deterioration of another important parameter. Routine engineering handles conflict situations through searching for the best compromise, or trade off between the different values the parameters can accept.
Altshuller found that the engineering conflicts can be indexed according to the type of parameters involved (39 common engineering parameters were defined). Examining numerous inventions made it possible to assign each conflict with a set of possible hints or strategies on how to approach the solution to the problem. Three types of hints are used: principles, standards and physical effects. The principles (40 of them) are high level strategies for solving the problem, standards (70 ) are more elaborated ideas based on past solutions, the knowledge base of physical effects (about 400) comprises a collection of physical, chemical and geometrical effects indexed according to the functions that each effect can carry out.
SIT differs from TRIZ in some fundamental aspects. The notion of overcoming conflicts is replaced with the application of the sufficient conditions. Reformulating the problem through the sufficient conditions generates a well defined and clear criterion for testing a candidate solution. Another important difference between the two methods is that SIT applies a minimal set of techniques, so that after some training the SIT process can become second nature to the problem solver. In TRIZ the user should always refer to a large external knowledge base.
SIT is used on an experimental basis in some Israeli hitech companies and in FORD Motor Company in the US (Sickafus 96). The method is taught as a full credited academic course in TelAviv University, and the National University of Singapore. Many creative solutions have been attained by SIT practitioners.
Section 2 presents an example of a technological problem, routine solutions and a creative solution. This example help in developing an intuitive impression of the sufficient conditions; then the formal definition of the sufficient conditions is presented in Section 3; Section 4 presents the empirical investigation of the sufficient conditions; Section 5 describes the SIT mechanism; Some examples of the application of SIT are presented in Section 6. Section 7 presents the conclusions and plans for future research.
THE SUFFICIENT CONDITIONS FOR CREATIVE SOLUTIONS
To present the sufficient conditions we start with an example of a technological problem. Some routine ideas to solve the problem, as well as one creative idea will be presented. A short analysis of the different solutions will show that the creative idea has unique properties not common to any of the routine ideas. These properties will then be generalized (yet compacted) and formulated as a set of two (jointly) sufficient conditions for creative solutions.
Example  Material Testing
A company develops materials that should withstand extremely harsh environmental conditions. Endurance tests are performed in a vessel, where samples are immersed in an acidic liquid at high temperature and pressure. The problem is that the vessel, being exposed to the acid at high temperature and pressure, does not withstand the conditions and has to be replaced frequently. The example is taken from Altshuller (85).
Routine Ideas. The following ideas are the most common ideas proposed by engineers confronted with the above problem
·1 To coat the vessel with a layer of protective material
·2 To mount another smaller vessel within the current one. That vessel will be constructed from cheaper material and will be replaced from time to time.
·3 To replace the vessel with one that is more durable in an acidic environment
An Inventive Solution. To drill a hole in the tested samples, and pour the acid into that hole. The samples will then be positioned in the vessel for the needed ambient conditions. The acid will not be in contact with the vessel, so it will not be damaged.
Analysis of the Solutions.The inventive solution exhibits two unique properties that do not coexist in any of the routine solutions. The first property is that no new element was added  the solution made use only of existing objects: vessel, acid, samples.
The second property is that a fundamental relation between two variables has qualitatively changed. The variables are the acid concentration and the frequency at which the vessel had to be replaced. While in the problem state, when the acid becomes more concentrated the vessel is more severely damaged and has to replaced more often, in the solution state the two variables are totally unrelated. Since there is no contact between the vessel and the acid, its concentration becomes irrelevant.
Examining the list of routine solutions reveals that none of these solutions have both properties. The first solution does not have the first property: a new element, the coating material was added. The second solution does not have both properties. The third solution does not satisfy the second property: even if the vessel is more durable, still the damage, albeit small, will be a function of the acid concentration. The first property will be generalized to become the closed world (CW) condition, while the second will be called the qualitative change in problem characteristic condition (QC). The theory asserts that any solution that satisfies these conditions simultaneously, will be deemed creative by field experts. Section 4 presents an experimental demonstration of the relation between the satisfaction of the conditions and evaluation of creativity by experts.
FORMAL EXPRESSION OF SUFFICIENT CONDITIONS
In this section we present a formal expression of the sufficient conditions. We will first present some notations and definitions, and then use them to formulate the sufficient conditions. The proposed system of notations is related to a situation in which a problem is described in terms of a given (existing) technological system that suffers from (known) undesired effects.
Notations and Definitions

the given system in the problem state (I for input) 

the system in the solution state (o for output) 

the neighborhood of a system (The collection of objects which are not an integral part of the system but can be found in the system’s proximity or have special affinity to that system) 

the collection of object types from which the system S is composed. Each object stands for the single technological concept that underlies its functioning in the system. 
UDE 
the collection of variables which contribute, directly or indirectly to the undesired effects, that appear in the problem description 

y is an increasing function of x, when all other variables remain constant. 

y is a decreasing function of x, when all other variables remain constant. 

the value of y is independent of the value of x 
If , is called a problem characteristic function.
Using these notations and definitions the two sufficient conditions can be expressed as follows:
The Closed World (CW) condition:
(1)
The Qualitative Change in Problem Characteristic (QC) condition:
for
(2)
The expression for the closed world condition means that no new object can be added to the system, unless it is a neighborhood object, but objects can be removed from the system. Since stand for object types and not the objects themselves, more objects of the same type are allowed to be introduced into the system (e.g. add more wheels to a car).
The expression for the qualitative change in problem characteristic condition means that a problem characteristic needs to change from an increasing function to either a decreasing or an unchanging function.
The sufficient conditions were developed through an empirical survey of numerous engineering problems and their corresponding routine and creative solutions. Once the conditions were extracted an explanation for the rational behind them may have been induced: Commonly routine design problem solving processes begin with an attempt to tune parameters, and when this fails to produce the desired results, engineers turn to searching alternative technological concepts. The QC condition makes parameter tuning ineffective, and the CW condition does not allow a replacement of existing concepts. Routine processes thus fail to produce the desired results, and the problem solver is forced to resort to more creative processes.
EMPIRICAL DEMONSTRATION OF THE RELATION BETWEEN CREATIVITY EVALUATION AND THE SATISFACTION OF THE CONDITIONS
The major purpose of the empirical test is to validate the sufficient conditions as appropriate conditions for creativity. The empirical test demonstrates the positive relation between the evaluation of creativity and the defined sufficient conditions. For the purpose of this study we conducted two preparatory stages. The first was designed to characterize the solutions to relevant engineering problems in terms of creativity scores. The second stage was designed to demonstrate that naive subjects can be instructed in a straightforward way, through written instructions, to apply the criteria of the sufficient conditions to the evaluation of solutions.
Given the results of the two stages, the third and final stage could be worked out by integrating the two sets of data presented in Table 1. One set of data (which is the outcome of Stage 1) consists in creativity scores. The second set of data (which is the outcome of Stage 2) consists in the evaluation of the conditions.
Stage 1: Creativity Evaluation
Subjects: A total of 196 engineers, 176 men and 20 women participated in the study. Their work experience ranged from 4 to 20 years.
Measures and procedures: Each subject was presented with one technological problem, selected randomly from a group of 20, and a description of 310 possible solutions for each problem. The problems were a randomly selected sample from a pretested collection of 50 technological problems, for each of which, a solution that satisfies the conditions is known to exist. The problems originate from different engineering domains such as mechanical engineering, electrical engineering and civil engineering. They were taken from different sources such as patent literature, the industry, and the authors’ own experience. We assume that the set of 50 problems is representative of the kind of problems that commonly crop up in engineering. Each set of solutions included at least one solution that satisfied the conditions as well as other solutions that are commonly offered by engineers (the solutions had been collected in numerous creative problem solving workshops conducted by the authors). The different solutions were presented to the subjects in a random order. The subjects were asked to rank each solution for its degree of inventiveness on a scale from 1 (not inventive at all) to 7 (very inventive).
Stage 2: Sufficient Conditions Evaluation
Subjects: Subjects were 3 engineers selected randomly out of 20 students who study for higher degrees at TelAviv University. Note that selecting 3 subjects for this experiment is not designed to extract statistical properties of any random distribution, but simply to enable cross checking of the subject’s evaluation. In principle the process of evaluating a solution for its satisfaction of the conditions is a deterministic process, although some margin for error should be allowed.
Measures and procedures: The subjects, who acted as judges in this experiment, were given a 20 page booklet that described the test procedure of the sufficient conditions. Each subject was then presented with the same set of 20 problems and their corresponding solutions that was used in Stage 1 of the experiment. The subjects were asked to test each solution for its compliance with the CW condition and the QC condition. A solution would be considered “satisfying the conditions in the judges view” if a majority of the judges considered it as satisfying the conditions. The results of Stage 2 are presented in Table 1 in the distribution of “J” signs.
Stage 3: Integrating the Results of Stages 1 and 2
Table 1 presents the integration of the results of Stage 1 and Stage 2. Each row represents a different problem and each cell in a row represents a possible solution. Each cell contains a numerical value which is the average score of the solution in addition to the (optional) symbols “*” and “J” to indicate that the particular solution was designated as satisfying the conditions by the authors and the judges correspondingly. In each row the solutions are ordered in a descending order of creativity evaluation score (Stage 1) for the ease of reading.
Table 1. THE RESULTS OF EXPERIMENTS 1 AND 2.
Prob. number 
No. of responses 
sol. 1 
sol. 2 
sol. 3 
sol. 4 
sol. 5 
sol. 6 
sol. 7 
1 
9 
5.1*J 
4.5J 
4.0 
2.7 
2.5 
2.3 
1.7 
2 
11 
6.4*J 
5.0 
4.9J 
4.9 
4.6 
3.8 
3.8 
3 
11 
5.9*J 
4.8J 
3.6 
3.0 
2.4 


4 
9 
5.6*J 
4 
3.2 
2.6 
2.4 


5 
9 
5.1*J 
4.4*J 
2.9 
1.6 



6 
10 
5.4*J 
3.7 
3.1 
3.0J 
2.4 
1.3 

7 
10 
6.0*J 
5.5 
4.8*J 
4.1J 
3.9 
3.1 

8 
11 
5.7*J 
5.3J 
5.2*J 
3.1J 
3.0 
1.7 
1.5 
9 
10 
6.5*J 
2.8 
2.3 
1.8 



10 
12 
4.8*J 
4.8*J 
4.1*J 
3.6 
3.2 
2.7 

11 
10 
5.3*J 
4.8*J 
3.0 
3.0 
2.8 
2.6 

12 
9 
5.9*J 
5.4 
2.9 
1.7 



13 
9 
4.8*J 
3.8J 
1.8 




14 
9 
6.2*J 
5.7*J 
4.3 
2.0 



15 
12 
5.2 
4.2*J 
1.6 




16 
9 
6.4*J 
4.9 
2.1 




17 
10 
5.6*J 
5.0J 
3.5 
3.4 



18 
9 
5.5*J 
5.3*J 
3.2 
2.5 
2.3 
2.2 

19 
9 
5.8*J 
3.7* 
3.5*J 
2.5 
1.5 


20 
9 
6.1*J 
5.1*J 
5.1 
1.7 



Analysis of the Results  Direct Observation
Scanning Table 1 reveals that, in all cases except for Problem 15, a solution that satisfies the conditions received the best creativity score. In all problems except for problems 6, 8, 13 and 19, the solutions that satisfy the conditions scored more than 4. In problems 6 , 8 and 13, however, those low rated solutions were not designated as satisfying the conditions by the authors.
The judges’ results were in accord with the authors’ in more than 90% of the cases. It is interesting to note that in most cases even when a particular solution was assigned ‘J’, but not ‘*’ this solution scored relatively high in creativity evaluation. This demonstrates the fact that the sufficient conditions are not very sensitive to interpretation errors.
Analysis of the Results  Pointbiserial Correlation Coefficient
The Pointbiserial correlation test is used to calculate the correlation between a two valued nonnumerical variable (called dummy variable) and a numerical continuous variable. A Pointbiserial correlation between the dichotomic dummy variable: 0 (= not satisfying the conditions), and 1 (= satisfying them), and creativity scores was computed. A value of 0.7 was obtained. The meaning of this value is that 49% of the variance in creativity scores can be explained by whether a solution satisfies or does not satisfy the conditions. This is considered a very high score in this type of experiments (involving human judgment). It means that with just these two values (of one variable) a very high portion of the variability is cleared. Figure 1 presents a graph showing creativity scores in relation to the conditions.
Figure 1. Creativity scores in relation to the conditions.
DESCRIPTION OF THE SIT MECHANISM
The SIT mechanism comprises three main steps: Problem reformulation through the sufficient conditions; selection of a general thinking strategy; selection and application of a relevant idea provoking technique. We will now describe each step in some detail. A more formal presentation of the SIT mechanism appears in Horowitz and Maimon, 97).
Problem Reformulation
At this stage the problem solver sets the target for the problem solving task using the two sufficient conditions. The CW condition is added to current constraints, while the QC condition changes the goal: instead of the initial (and natural) requirement to decrease the level of an undesired effect, the problem solver is guided to qualitatively change a mathematical relation between any two problem related variables (problem characteristics).
Technically at this stage the user forms a list of system objects, a list of system neighborhood objects, and a list of problem characteristic variables. The problem solving task is defined as follows: Find a solution in which at least one of the defined increasing functions will become decreasing or unchanging subject to the constraint that the solution will incorporate only elements of the given system and its neighborhood that appear in the relevant lists.
Strategy Selection
The framework of the sufficient conditions naturally gives rise to two thinking strategies. A candidate solution is composed of three elements: the desired physical end state  deduced from the QC condition, the objects to be modified , and the required modification. The CW condition confines the objects to be modified only to existing ones, and thus significantly narrows the search space. Two scenarios are possible at this stage:
1. The problem solver can deduce a required physical end state from the QC condition. This situation commonly occurs when the desired end state can be achieved through a simple physical operation, that will not interfere with other operations required from the system.
2. The problem solver cannot conceive a desired physical end state, or the state he can think of contradicts other fundamental requirements from the system.
Theses two scenarios define the two possible strategies. Following the first strategy, the problem solver first formulates a conceptual solution: a simple operation, that once added to the system, the QC condition is guaranteed to be satisfied. He then proceeds to find an existing object that will carry out the desired operation. For example in the course of the solution of the material testing problem, the problem solver can think of the idea to physically separate the acidic liquid from the vessel  an idea that guarantees the satisfaction of the QC condition. He then selects an existing object: the tested samples to carry out this operation. This strategy is called the extension strategy to indicate the fact that the system is temporarily extended through the addition of an imaginary object that will carry out the new operation.
If the problem solver reaches at this point the second situation, that is he cannot conceive of a desired end state that would guarantee the satisfaction of the QC condition, he can follow a different strategy: through a trial and error process he tries different possible modifications to existing objects until at some point hopefully he hits a state where the QC condition is satisfied (the CW condition is guaranteed to be satisfied since none of the tried modifications violates it). This strategy is called the restructuring strategy to indicate the fact that in the trial and error process the problem solver changes the structure of existing objects and their organization.
The problem solver is guided to select the extension strategy if he can conceive a conceptual solution, and to select the restructuring strategy otherwise. The actual significance of selecting a thinking strategy lies in the application of a different set of idea provoking techniques for each strategy. If the extension strategy was selected the user is directed to apply either unification or multiplication, if the restructuring strategy is selected the user is directed to apply either division, or increasing variability. The extension techniques help the problem solver identify an existing object that will carry out the new operation, while the restructuring techniques help him increase the degrees of freedom of possible changes to the system.
Idea Provoking Techniques
Idea provoking are the final stage of the method. Their main role is to free the problem solver from fixated mental states.
The Unification Technique. the unification technique helps the problem solver identify a system or neighborhood objects that will carry out the operation defined in the conceptual solution. Applying the technique is a four step process:
1. Formulate the needed operation
2. Form a list of all main system and neighborhood objects
3. Select an object from the list and complete the following sentence: The selected object will carry out the operation
4. Determine the necessary modifications of the selected object, so that it can carry out the desired operation
Example  the four steps applied to the material testing problem:
1. The needed operation: to separate the acidic liquid from the vessel
2. A list of objects: vessel, samples, acidic liquid
3. Selected object: samples. The samples will separate the acidic liquid from the vessel
4. required modification: the shape of the samples will change so that it can contain liquid
The Multiplication Technique. The purpose of this technique and its first 2 steps (out of four) are identical to the Unification technique. The last two steps are listed below:
1. Select an object from the list and complete the following sentence: The selected object will multiplied. The new copy (or copies) of the of this object will carry out the operation
2. Determine the necessary modifications of the new copy (or copies) of the selected object, so that it can carry out the desired operation.
The Division Technique. Being a restructuring strategy technique the purpose of the division technique is to help the user identify new degrees of freedom for modifying and reorganizing system objects. It is a 3 step process:
1. Form a list of system objects
2. Select an object from the list and complete the following sentence: The object will be divided to its more basic elements/to smaller parts of the same part/in a random way [select one option from the three]
3. Search for meaning: try to use the new degrees of freedom to create a state in which the QC condition is satisfied: different parts in different locations, different order of parts etc.
The Increasing Variability Technique. This is another important technique to aid the problem solver in the creation of new degrees of freedom and new ways to solve the problem. It is a 4 step process:
1. Form a list of system objects
2. Select an object
3. Select two parameters W and Z that are currently not related, that is W is not a function of Z (a new degree of freedom will be the type of relation between them).
4. Search for meaning: try to use the new degrees of freedom to create a state in which the QC condition is satisfied
EXAMPLES FOR THE APPLICATON OF SIT
This section will demonstrate how SIT is used to find creative solutions to technological problems. Each example will include problem description, a set of routine solutions (most common responses of engineers), detailed description of how SIT is used to solve the problem, and a description of a creative solution  the output of the SIT process. It is important to note that common SIT applications consist in many trial and error processes (different strategies, different techniques, different application of the techniques), however the description below does not reflect that important nature of SIT application. For sake of brevity we describe in each example only the direct path to the solution.
Example No. 1  Corn in a Pipe
A curved steel pipe is one of the components of a corn grain processing plant. The pipe's function is to conduct the flow of air with the corn grains. The problem is the grains’ impact on the pipe at the bend, which erodes the pipe wall. The air speed cannot be reduced since this will reduce the plant's capacity. The problem first appeared in Altshuller (85).
Figure 2. The corn in the pipe problem
Routine ideas:
1. To strengthen the pipe at the erosion zone (change the material, make it thicker)
2. To make the curved part of the pipe from a different piece that can be easily replaced
3. To coat the pipe with a protective layer that will be replaced from time to time
SIT step 1  problem reformulation:
List of UDE parameters: cost of production, erosion rate, grain flux, grain hardness
List of system and neighborhood objects: pipe, grain, air flow
The reformulated problem: make erosion rate unrelated to/decreasing function of grain flux, don’t add any new object to: pipe, grain and air flow.
SIT Step 2  strategy selection:
Since a conceptual solution can be conceived  to separate the grains from the pipethe extension strategy is selected.
SIT Step 3  select and apply an idea provoking technique (unification or multiplication)
Unification was selected, application of the technique:
1. Formulate the needed operation: to separate the grains from the pipe
2. Form a list of all main system and neighborhood objects: grains, pipe, air flow
3. Select an object from the list [grain] and complete the following sentence: The grain will separate the grain from the pipe
4. Determine the necessary modifications of the selected object, so that it can carry out the desired operation: grains should stick to the curved part of the pipe
The solution: The geometry of the curved area of the pipe will change as to create a pocket that will enable the grains to accumulate there. This will protect the pipe from the damage of grains’ impact (see Figure 3).
Figure 3. The solution to the corn in a pipe problem.
Example 2: Derailing detection device
The braking system of trains includes a pipe that passes along the train, in which the air is at a pressure of 5 atmospheres. When the pressure drops, the train stops. Under emergency conditions (such as derailing), the air must be released very quickly. To ensure fast enough release of the air, it should exit through an opening that is at least 10 cm2. During normal operating conditions, this opening should be closed with a stopper. The stopper should be released by the air pressure itself.
A new derailing detector has been developed. The idea is that in normal operation, the stopper is to held in place by the derailing detector, and when derailing occurs the detector stops exerting force on the stopper and it is released. The problem is that the derailing detector can exert only 0.5 Kgf, not enough to balance the 50 Kgf applied by the internal pressure.
Figure 4: The derailing detector system.
Routine ideas:
1. To use a lever
2. To add more derailing detectors, each will support a smaller stopper
3. To squeeze the stopper in its place so that friction will carry out some of the load
SIT step 1  problem reformulation:
List of UDE parameters: Probability of false alarm, probability of premature stopper opening, load on the derailing detector, air pressure, stopper area
List of system and neighborhood objects: pipe, air, stopper, derailing detector
The reformulated problem: make load on the derailing detector unrelated to/decreasing function of air pressure, don’t add any new object to: pipe, air, stopper, derailing detector.
SIT Step 2  strategy selection:
Since a conceptual solution can be conceived  to exert on the stopper a force that is identical and in opposite direction to the force exerted by air pressure, the extension strategy was selected.
SIT Step 3  select and apply an idea provoking technique (unification or multiplication)
Multiplication was selected, application of the technique:
1. Formulate the needed operation: to exert on the stopper a force that is identical and in opposite direction to the force exerted by air pressure
2. Form a list of all main system and neighborhood objects: pipe, air, stopper, derailing detector
3. Select an object from the list [stopper] and complete the following sentence: The stopper will be multiplied. The new copy (or copies) of this object will exert on the stopper a force that is identical and in opposite direction to the force exerted by air pressure
4. Determine the necessary modifications of the new copies of the selected object, so that it can carry out the desired operation: The new stopper should be slightly smaller than the original one so that the derailing detector will still have to carry some load .
The solution: The new stopper will be mounted exactly above the original one, they will be connected through a thin wire (see figure 5).
Figure 5: The solution to the derailing detector problem
Example No. 3  The Tumor Problem
Suppose you are a doctor faced with a patient who has a malignant, inoperable tumor in his stomach. Unless the tumor is destroyed the patient will die. There is a ray that can be used to destroy the tumor. If the rays are directed at the tumor at sufficiently high intensity, the tumor will be destroyed. Unfortunately, at this intensity, the healthy tissue that the rays pass through on the way to the tumor, will also be destroyed. At lower intensities, the rays are harmless to the healthy tissue but they will not affect the tumor.
Routine ideas:
In this problem the most common ideas either violate problem definition (for example to try to operate although the problem text explicitly states that the tumor is inoperable) or suggest alternative treatment such as chemotherapy.
SIT step 1  problem reformulation:
List of UDE parameters: Probability of patient's death, damage to healthy tissues, rays intensity
List of system and neighborhood objects: rays, tumor, healthy tissues
The reformulated problem: make damage to healthy tissues unrelated to/decreasing function of rays intensity, don’t add any new object to rays, tumor, healthy tissues.
SIT Step 2  strategy selection:
Since a conceptual solution cannot be conceived the restructuring strategy is selected.
SIT Step 3  select and apply an idea provoking technique (division or increasing variability)
Division was selected, application of the technique:
1. Form a list of system objects: rays (this is the only system object)
2. Select an object from the list [rays] and complete the following sentence: The rays will be divided to smaller parts of the same type
3. Search for meaning: try to use the new degrees of freedom to create a state in which the QC condition is satisfied (different parts in different locations, different order of parts etc.): The smaller rays will be directed at the tumor from different angles.
The solution: To direct a few weak beams at the tumor from different angles, so that they converge at the tumor and develop sufficient intensity there to destroy the tumor. The QC condition was satisfied since it is possible to increase the ray intensity at the tumor by adding more weak rays without affecting healthy tissues through which the rays pass.
Example No. 4  Solid Fuel Rocket Engine
One of the problems that designers of solidfuel rocket engines faced was the necessity of achieving a constant thrust from the engines. The solidfuel rocket engine has the shape of a hollow cylinder burning in an internal envelope. The problem with such geometry is that the thrust is not constant owing to a change in the area of the internal envelope (the radius increases). When the internal combustion area increases, the thrust increases.
Figure 6. Side view of a rocket engine.
Routine ideas:
1. A new parametric design: The dimensions of the cylinder are changed so that it will become longer and narrower. These changes maintain its total volume and combustion area, but since the difference between initial and final radius is smaller, the variance is smaller.
2. “Cigar burning”  a cylinder burning in its base
SIT step 1  problem reformulation:
List of UDE parameters: Energy waste, uneven thrust, thrust increase, burning area increase, perimeter increase
List of system and neighborhood objects: solid fuel, rocket, thrust
The reformulated problem: make burning area unrelated to/decreasing function of perimeter, do not add any new object to solid fuel, rocket, thrust.
SIT Step 2  strategy selection:
Since a conceptual solution cannot be conceived the restructuring strategy is selected.
SIT Step 3  select and apply an idea provoking technique (division or increasing variability)
Increasing variability was selected, application of the technique:
1. Form a list of system objects: solid fuel, rocket, thrust
2. Select an object: solid fuel
3. Select two parameters W and Z that are currently not related, that is W is not a function of Z (a new degree of freedom will be the type of relation between them): Z  cross section shape; W  combustion progression.
4. Search for meaning: try to use the new degrees of freedom to create a state in which the QC condition is satisfied: The shape of the cross section will change through the combustion progression from a very complicated winding shape to a pure circle.
The solution: the shape of the crosssection is such that it maintains a constant perimeter while combustion progresses. The crosssection changes from a complex shape to a simple circle, thus, although the average radius increases, the perimeter remains constant. Figure 7 demonstrates the idea. This solution preserves the initial concept of a hollow shape burning in the internal envelope thus complying with the closed world condition. Since the variance in thrust is constantly zero, totally independent of the difference between initial and final radius, the solution satisfies the QC condition as well. Note that, at its time, this solution was a breakthrough in solid fuel engines. In our workshops we see students using the sufficient conditions finding this solution quite quickly.
Figure 7. The new inner envelope as combustion progresses
CONCLUSIONS AND PLANS FOR FUTURE RESEARCH
The paper presented a set of objective sufficient conditions for creative solutions, and a threestep method that structures the search toward conditions satisfying solutions. The application of the sufficient conditions to a specific problem modifies the given problem definition: The QC condition changes the goal of the search and the CW condition confines the search space. By applying the sufficient conditions, the creative problem solver actually solves a different problem than his noncreative companion, an important factor in finding a different, sometimes surprising solution.
The validity of the sufficient conditions was demonstrated through the empirical study presented in Section 4. It was shown that conditions satisfying solutions score high in field experts’ evaluation of creativity. We are now in the midst of an ongoing research aimed at empirically demonstrating SIT effectiveness in solving creatively engineering problems. We also study the deeper cognitive changes in SIT trainees, using the method of individual meaning profile assessment developed by Kreitler and Kreitler (68).
Intermediate results indicate that considerable improvement is achieved in the subjects’ ability to solve creatively engineering problems. The rate of those who succeed in finding conditions satisfying solutions increases from an average of 8% before the course to more than 30% after the course. The cognitive test shows some interesting changes in cognitive styles before and after SIT training.
Main Differences between SIT and TRIZ
Now that the SIT mechanism has been unfolded lets summarize the main differences between SIT and its predecessor TRIZ:
·1 SIT uses a minimal set of techniques: 5 idea provocative techniques in contrast to hundreds techniques used in TRIZ. This SIT feature shortens the path for a desired solution.
·2 SIT mechanism is tighter in keeping the problem solver within the framework of truly creative ideas. The reason for this is that the sufficient conditions: the Closed World condition and the Qualitative Change in Problem Characteristic condition that replaces TRIZ’s requirement for overcoming a contradiction pose easytotest requirements on candidate ideas. TRIZ literature contains many illustrative examples of solutions that are said to have overcome a contradiction, but do not satisfy the sufficient conditions.
·3 SIT training is different than TRIZ training. Due to the compact and streamlined structure of the method the SIT trainee is exercising the same techniques time and again on different problems until they become his second nature. When using the method in real life the problem solver should not refer to any external knowledge base, but retrieve the tools from his own long term memory.
In spite of these differences the two approached are complementary. At times searching a large database of solution principles, as suggested in TRIZ, may prove useful. SIT training, however, is a good way to train an individual in drawing meaning out of TRIZ’s ideas bank.
Future research
In the future we intend to develop a computerized environment that will support SIT learning, exercising, application and team work. The system will also support the creation and maintenance of a data base of examples of past creative solution, which will store not only the solution, but also the process that led to the solution (including dead ends). A data base of structured solution processes may become an important asset of organizations.
References
Allen M. S., Morphological Creativity, PrenticeHall New Jersey, 1962
Altshuller G. S., Creativity as an Exact Science, New York, Gordon and Breach, 1985
Dasgupta, S., Creativity in Invention and Design, Cambridge University Press, 1994
de Bono E., Serious Creativity, Fontana, 1992
Fey R., E. I. Rivin, Vertkin I. M. “Application of the Theory of Inventive Problem Solving to Design and Manufacturing Systems”, Annals of the CIRP, 43, January, 1994
Gordon W .J. , Synectics, Harper and Brothers, New York , 1961
Hennessey A. and Amabile T. M., “The Conditions of Creativity”, in The Nature of Creativity, Robert J. Sterenberg (ed.), pp. 1138, Cambridge University Press, 1988.
Horowitz R., Maimon O., “SIT (Structured Inventive Thinking)  A Method for Enhancing Creative Problem Solving in Engineering”, submitted for publication in IEEE Systems Man and Cybernetics, 1997
Kreitler H., Kreitler S, “dimensions of meaning and their measurement”, Psychological Reports 23, pp. 13071329, 1968
Maimon, O., Horowitz, R. “Sufficient Conditions for Design Inventions”, submitted for publication in IEEE Systems Man and Cybernetics.
Malmqvist J., Axelsson R., Johansson M., “A Comparative Analysis of the Theory of Inventive Problem Solving and the Systematic Approach of Pahl and Beitz”, Proceedings of the 1996 ASME Design Engineering Technical Conference, 1996
Osborn, A. F., Applied Imagination, New York: Charles Scribner's Sons, 1959.
Sickafus E. N., “Structured Inventive Thinking”, The Industrial Physicist, March 1996
Sushkov V., Mars N. J., Wognum P. M., “Introduction to TIPS: a theory for creative design”, Artificial Intelligence in Engineering, 9, 1995
