Date: 11/02/07 (Algorithms)    Keywords: technology

I pose the question: Is finding the ideal mate is an intractable problem or are we just not dedicating sufficient computational resources to the problem presently?

The ability to identify and assess every person in the world to determine which one is best suited for one's self is presently beyond the informational and computational power of any one person. Can present or near-future computational power be successfully harnessed to increase the success of lifelong intimate relationships?

Let a 'match' be a pairing of two individuals for the purpose of being lifelong romantic partners.
Let 'suitability' be a measure of how suitable person A is for person B in the event of A being matched with B.
Let an 'ideal match' be one where the measure of the sum of A suitable for B and B suitable for A is maximized.

The methods that dominated courtship over the last century require the relatively slow biological computational processes within our human bodies in general and brains in particular--we must meet and get to know other people. Even then, the accuracy of such processes may be questioned, as the time devoted to courting or 'assessing' any one particular 'match' may lead to an irrational 'escalation of commitment' and a premature, if not ill-advised coupling may ensue.

Over the past few decades, online interactions in general and online dating services in particular have promised a more streamlined approach to culling through the candidates to find those matches that may be more ideal, allowing people to focus on assessing those potentials that are most promising. eHarmony.com, for example, presents its members with extensive personality and interests inventory questions and claims to 'match' its members based upon 'algorithms'.

I have only recently began my study (read: 'use') of these online dating services however my initial assessment is that services like these and their future incarnations may be able to at least find some of us matches that are 'near ideal'.

However, I suspect that a related problem--that of maximizing the total 'ideal-ness sum' for all matches--is ultimately NP-hard. Clearly as we match more of us up, there will be a smaller pool from which to find 'near ideal' matches for the rest of us. In fact, some matches may have negative suitability, leading to the intuitive result that some of us just will not be matched, once the pool of remaining candidates becomes relatively small. I believe our friends in the Economics Department will find interest in this problem as well.

This leads me to four conclusions:
1. Whatever the methods of courtship available to us, whatever the technology employable, the result remains the same: find someone suitable before it is too late. The childhood exercise of 'musical chairs' clearly illustrates this point.
2. Technological assisted matching may provide some of us an edge in finding a 'near deal' match, or 'more ideal' match. However, we do not all share equal access to these resources, meaning that finding a suitable match is more likely for those of us with privilege.
3. Each 'match' commitment makes the pool of potentials smaller for the remaining unmatched individuals. This benefits no one but the pair involved in the match and likely harms the potential of some of the remaining unmatched individuals as the pair effectively removes themselves from the pool.
4. I understand why my mother gave up on men after several failed relationships. Those men remaining unmatched in their fifties are just not likely good matches for anyone. Clearly, if she had been more privileged as a young woman, she would have had a larger pool of suiters from which to choose, and would have had a better chance to find a 'near ideal' match.

Source: http://community.livejournal.com/algorithms/94578.html