Doesn't the "NP=P? Problem" essentially boil down to the question:
"What 'network characteristics' can be observed, measured and encoded in an algorithm, apriori, such that, at run time, we:
(a) pre-empt a hopeless combinatorial explosion AND
(b) mitigate processing time with satisfactory, if not optimal, results?"
If so, wouldn't it be most fruitful to use some combination of linear programming and similar stochastic methods? If not, what am I missing here?
Slashdot Mirror
Doesn't the "NP=P? Problem" essentially boil down to the question: "What 'network characteristics' can be observed, measured and encoded in an algorithm, apriori, such that, at run time, we: (a) pre-empt a hopeless combinatorial explosion AND (b) mitigate processing time with satisfactory, if not optimal, results?" If so, wouldn't it be most fruitful to use some combination of linear programming and similar stochastic methods? If not, what am I missing here?