Approximation Algorithm
An approximation algorithm is a strategy to find near-optimal solutions to optimization
problems, particularly those classified as NP-hard, within a reasonable amount of time. These
algorithms are useful when finding the exact solution is computationally infeasible.
Key Features
1. Efficiency:
o Approximation algorithms run in polynomial time, making them suitable for
large input sizes.
2. Approximation Ratio:
o Measures how close the approximation solution is to the optimal solution.
o If CapproxC_{\text{approx}}Capprox is the cost of the solution from the
approximation algorithm and CoptC_{\text{opt}}Copt is the cost of the
optimal solution: Approximation Ratio (R)=max(CapproxCopt,CoptCapprox)\
text{Approximation Ratio (R)} = \max \left(\frac{C_{\text{approx}}}{C_{\
text{opt}}}, \frac{C_{\text{opt}}}{C_{\text{approx}}}\
right)Approximation Ratio (R)=max(CoptCapprox,CapproxCopt)
Applications
Traveling Salesman Problem (TSP).
Vertex Cover.
Knapsack Problem.
Set Cover.
Limitations of Approximation Algorithms
1. Approximation Quality:
o The solution may not be close enough to the optimal, especially for problems
with high approximation ratios.
2. Not Always Applicable:
o For some problems, finding even an approximate solution can be as hard as
finding the exact solution.
3. No Guarantee for Specific Inputs:
o Performance depends on the problem instance. Some instances might yield
poor approximations.
4. Complexity of Implementation:
o Approximation algorithms can still be intricate and difficult to implement.
5. Trade-off Between Accuracy and Efficiency:
o Higher accuracy often comes at the cost of increased computational
complexity.
Example 1: Vertex Cover
Problem:
Given a graph G=(V,E)G = (V, E)G=(V,E), find the smallest subset of vertices such that
every edge in EEE has at least one endpoint in the subset.
Approximation Algorithm:
1. Start with an empty set CCC.
2. While there are edges left in the graph:
o Pick an arbitrary edge (u,v)(u, v)(u,v).
o Add both uuu and vvv to CCC.
o Remove all edges incident to uuu and vvv.
3. Return CCC.
Analysis:
Approximation Ratio: 222 (the solution is at most twice the optimal size).
Example:
Graph: V={1,2,3,4},E={(1,2),(2,3),(3,4)}V = \{1, 2, 3, 4\}, E = \{(1, 2), (2, 3), (3,
4)\}V={1,2,3,4},E={(1,2),(2,3),(3,4)}.
Approximation Solution: {2,3}\{2, 3\}{2,3} (covers all edges).
Optimal Solution: {2,3}\{2, 3\}{2,3} (in this case, the approximation is exact).
Example 2: Traveling Salesman Problem (TSP)
Problem:
Find the shortest possible route that visits every city exactly once and returns to the starting
city.
Approximation Algorithm:
Use Minimum Spanning Tree (MST):
1. Compute the MST of the graph.
2. Perform a preorder traversal of the MST.
3. Return the traversal as the approximate tour.
Approximation Ratio:
For metric TSP (distances satisfy the triangle inequality), the approximation ratio is
222.
Example:
Graph: A→B:10,B→C:15,C→A:20A \to B: 10, B \to C: 15, C \to A:
20A→B:10,B→C:15,C→A:20.
MST: A−BA-BA−B, B−CB-CB−C.
Approximate Tour: A→B→C→AA \to B \to C \to AA→B→C→A, cost = 45.
Optimal Tour: A→B→C→AA \to B \to C \to AA→B→C→A, cost = 45 (optimal and
approximate are the same).
Conclusion
Approximation algorithms provide a trade-off between computational efficiency and
solution quality. They are practical for NP-hard problems, but their limitations, such as
suboptimal results and variability with input, must be carefully considered.
