Reduction: RUDRATA CYCLE → TSP(Travel Salesman)
Given a graph , construct the following instance of the TSP:
- The set of cities is the same as
- The distance between cities and is 1 if is an edge of and otherwise, for some to be determined.
- The budget of TSP
Easy to see that if has a Rudrata cycle, then the same cycle is also a tour within the budget of the TSP instance
Conversely, if has no Rudrata cycle, then there is no solution: cheapest possible TSP tour has cost at least (must use at least one edge of length , and total length of all others is at least ).
We introduced the parameter because by varying it we can obtain interesting results
-
- all distances are either 1 or 2, and instance of TSP satisfies the triangle inequality.
- This is a special case of TSP which is of practical importance, because it can be efficiently approximated.
- is large
- resulting instance of the TSP may not satisfy the triangle inequality
- But either it has a solution of cost or less, or all solutions have cost at least (which now can be arbitrarily larger than ), nothing in-between!
- This is called an important gap property such that unless , no approximation algorithm is possible for this particular instance