Sie sind auf Seite 1von 7

Clarke-Wright Algorithm Basic Information For a bakery that delivers to 2 concession stands, vehicles could be routed in two different

t ways.

Concession 1

6 miles Concession 2 5 miles 0 Bakery

5 miles

First Option: Bakery to Concession 1 (5 miles), Concession 1 to Bakery (5 miles), Bakery to Concession 2 (5 miles), Concession 2 to Bakery (5 miles) Total distance = 2D01 + 2D02 = (2 * 5) + (2 * 5) = 20 miles Second Option: Bakery to Concession 1, Concession 1 to Concession 2, Concession 2 to Bakery Total savings = = = = D0i + D0j - Dij D01 + D02 - D12 5 + 5- 6 4

If the problem is more complicated and we now have 4 concessions, how should we determine the optimal distribution route which minimizes the distance traveled? 0 = Bakery 1 = Concession stand 1 2 = Concession stand 2 3 = Concession stand 3 4 = Concession stand 4

6 2 12 14 9 8 13 8 4 3 6

Step 1: Develop the basic half-matrix reflecting the distances between the bakery and the concession stands. 1 0 8 1 2 14 12 2 3 8 9 6 3 4 7 13 8 6 4

Step 2: Calculate the net savings if we travel between concession stands before returning to the bakery (remember that we use the formula D0i + D0j - Dij when determining the savings in distance). We will examine the following relationships.
a

D0i + D0j - Dij = D01 + D02 - D12 = 8 + 14 12 = 10

D0i + D0j - Dij = D02 + D03 - D23 = 14 + 8 6 = 16

1 0 1

2 10a 2

3 7 16b 3

4 2 13 9 4 3

Step 3: Determine the value for T. T identifies whether the trip is roundtrip or one-way T = 2 when the vehicle travels from its point of origin (o) to another point (i) and then back to the point of origin. T = 1 when the vehicle travels one way directly between two locations T = 0 when a vehicle does not travel directly between two particular Examples: For the trip [bakery(0) concession 1(1) bakery (0)] the T = 2. For the trip [concession 1 (1) concession 2 (2)] the T = 1. If no trip has been made between pairs of designations, the T = 0. Step 4: Run iterations until you find the best fit. Initial Solution

2 1 0

3 4

1 0 2 1

2 2 10 2

3 2 7 16 3

4 2 2 13 9 4 Distance = 8+8+14+14+8+8+7+7=74 miles

Step 4a: Look for the highest non-underlined distance. If it meets the following conditions choose it: T0i and T0j must be greater than zero Locations i and j are not already on the same route or loop Route 2,3 or 3,2 is 16 miles First Iteration

2 1 0

3 4

1 0 2 1

2 1 10 2

3 1 7 1 3

4 2 2 13 9 4 Distance = 8+8+14+6+8+7+7=58 miles

Step 4b: Look for the highest non-underlined distance. If it meets the following conditions choose it: T0i and T0j must be greater than zero Locations i and j are not already on the same route or loop Route 2,4 or 4,2 is 13 miles

Second Iteration

2 1 0

3 4

1 0 2 1

2 0 10 2

3 1 7 1 3

4 1 2 1 9 4 Distance = 8+8+8+6+8+7 = 45 miles

Step 4c: Look for the highest non-underlined distance. If it meets the following conditions choose it: T0i and T0j must be greater than zero Locations i and j are not already on the same route or loop Route 1,2 or 2,1 is 10 miles. However, this route violates the first condition. T01 is 2 but T02 is 0. The next largest distance is route 3,4 or 4,3 with a distance of 9 miles. However, this route violates the second condition. Concessions 3 and 4 are on the same route. The next largest distance is route 1,3 or 3,1 with a distance of 7 miles. Concession 1 and 3 are not currently on the same route and neither T01 or T03 is 0.

Third Iteration

2 1 0

3 4

Distance = 8+9+6+8+7 = 38 miles

1 0 2 1

2 0 10 2

3 1 1 1 3

4 1 2 1 9 4

Step 4d: Look for the highest non-underlined (or discarded) distance. If it meets the following conditions choose it: T0i and T0j must be greater than zero Locations i and j are not already on the same route or loop Route 1,4 or 4,1 with a distance of 2 miles is the only remaining route. However, this route violates the second condition. Concessions 1 and 4 are already on the same route. So you know you are done with the problem.

Das könnte Ihnen auch gefallen