This is an offshoot of another discussion: Can "expected future happiness" be quantified, or at least ranked among alternative futures?, where both respondents brought up the concept of the undersireabilty of inducing happiness artificially.
An interesting paper was brought up by the second commenter that stated that people are bad at predicting the intensity and duration of future experience, but not the general sense (positive/negative) of the experience. However, people are also not very good at guaging their adaptability to new situations.
My question is this: If you could be hooked up to a machine that would allow you to be as happy as you desired (via simulating scenarios and experiencse you need to feel as such) while also ensuring that your physical body is well taken care of and safe, then why is this not a way of optimizing well being? Even if we initially feel that we would not like such "fake" experiences, it seems that the emotional value of the experiences themselves, qua experiences, are positive and desireable, its just the epistemological packaging of "not really real" that is not desired.
Therefore, if someone were involuntarily hooked up to such a machine, what would their likely emotional trajectory? The article liked above and some introspection would suggest that while they would initially be angry/depressed, they would adapt to their new reality and seek ways to exploit its fluidity. A cognitive "reframing" would likely take place to reduce the dissonance between their inescapable situation and their feelings towards it, leading them to embrace it and the hedonistic optimization it allows.
Therefore, in the end, after an adjustment period, they appear to have a MUCH higher probabilty of achiveing MUCH higher happinness (on balance) by being hooked up to the machine. At least, that is my opinion (BTW: if it wasn't obvious, I am NOT a psychologist, just an interested layperson)