Importance of defining a problem
|Image Credits: http://openclipart.org/user-detail/technoargia|
Alternatively you could define the problem within a boundary. The same problem could be defined in other ways like - 'I want to reduce traffic jams in a city without building new roads' or 'I want to reduce traffic jams in a city without increasing public transport' etc. This way the problem and the constraints are known and therefore any further ideation, solution is more focussed. All this probably increases the chances of a solution being successful.
Another way to define a problem is as a contradiction. An example would be - 'I want to create a glass that is opaque'. Sounds impossible? But that's exactly how 3M defined the problem and came up with a glass that can be both transparent and opaque (an electric charge is applied to turn on / off the transparence). Examples at - http://electricglasswall.com/index.php?option=com_content&view=article&id=60&Itemid=27#g_60_0.
I have picked up these approaches from the TRIZ (Theory of Inventive Problem Solving) framework. There is no one approach, I suppose it depends on the context. What approach do you use?