Sei sulla pagina 1di 14

Assignment Problem

A typical assignment problem, presented in the classic manner, is shown in Fig. 12.
Here there are five machines to be assigned to five jobs. The numbers in the matrix
indicate the cost of doing each job with each machine. Jobs with costs of M are
disallowed assignments. The problem is to find the minimum cost matching of
machines to jobs.

Matrix model of the assignment problem.

The network model is in Fig. 13. It is very similar to the transportation model
except the external flows are all +1 or -1. The only relevant parameter for the
assignment model is arc cost (not shown in the figure for clarity) ; all other
parameters should be set to default values. The assignment network also has the
bipartite structure.

Network model of the assignment problem.

The solution to the assignment problem as shown in Fig. 14 has a total flow of 1 in
every column and row, and is the assignment that minimizes total cost.

Solution to the assignment Problem


Shortest Path Problem
This problem uses a general network structure where only the arc cost is relevant.
A typical case is shown in Fig. 15. The length of a path is the sum of the arc costs
along the path. The problem is to find the shortest path from some specified node
to some other node or perhaps to all other nodes. The latter problem is called the
shortest path tree problem because the collection of all shortest paths from a
2

specified node forms a graph structure called a tree. Since it is not much more
difficult to find the paths to all nodes than it is to find the path to one node, the
shortest path tree problem is usually solved.

Network model for the shortest path problem.


The network flow equivalent to the shortest path tree problem is formed by
equating arc distance to arc cost, assigning a fixed external flow of m - 1 (m is the
number of nodes) to the source node, and assigning fixed external flows of -1 to
every other node. This is illustrated in Fig. 15. When solving this flow problem, the
computer will assign flow from the source to each node by the least cost path,
since there are no bounds on arc flows. The shortest path tree will consist of those
arcs with nonzero flow in the optimum solution. The solution to the example is
shown in Fig. 16.

Solution of the shortest path problem.


Maximum Flow Problem
The maximum flow problem is again structured on a network; but here the arc
capacities, or upper bounds, are the only relevant parameters. The problem is to
find the maximum flow possible from some given source node to a given sink
node. A network model is in Fig. 17. All arc costs are zero, but the cost on the arc
leaving the sink is set to -1. Since the goal of the optimization is to minimize cost,
the maximum flow possible is delivered to the sink node.

Network model for the maximum flow problem.


The solution to the example is in Fig. 18. The maximum flow from node 1 to node
8 is 30 and the flows that yield this flow are shown on the figure. The heavy arcs
on the figure are called the minimal cut. These arcs are the bottlenecks that are
restricting the maximum flow. The fact that the sum of the capacities of the arcs on
the minimal cut equals the maximum flow is a famous theorem of network theory
called the max flow min cut theorem. The arcs on the minimum cut can be
identified using sensitivity analysis.

Solution. Maximum flow = 30.


Other Examples
A Transportation Problem with Costs of Manufacture and Revenue at Sale
A company has two warehouses and four customers. The shipping costs between
warehouses and customers are shown in the table below. Minimum and
maximum shipments and unit manufacturing costs are shown for each plant.
Minimum and maximum receipts, and unit revenues are shown for each
customer. Our problem is to model the problem of finding the optimum shipping
5

pattern with a network flow model.

The network model in Fig. 19 resembles that of a transportation problem except we


have added arcs to carry the extra supplies and demands allowed.

Network Model
Multiperiod Operation
The company needs to establish a shipping schedule for the next two months. The
demands for each customer are 15 units in the first month and 20 units in the
second month. These demands must be met. The warehouses are really
manufacturing plants where the products are made. Plant A has a manufacturing
capacity of 30 in each month, while plant B has a capacity of 50 in each month. In
6

the first month the cost of manufacture at A is $8 per unit and the cost of
manufacture at B is $10 per unit. In the second month the cost of manufacture is
$9 at both plants. Products can be stored at the customer sites from one month to
the next. The storage cost is $1 per unit. Products cannot be stored at the plants.
Shipping costs are as given in the previous example except that the shipping
company is giving a discount of $1 per unit on all routes during the first month.
The goal is to minimize total production, shipping and inventory costs over the
two months. Not all production capacity need be utilized by the solution.
The network model for this case is shown in Fig. 20. A transportation network
represents each period. Arcs that go from one period to the next represent
inventories. With this construction, the size of the network is proportional to the
number of periods. This same approach is useful for a variety of multiperiod
situations.Figure 20. Two period network model.

Two period network model.


Transformation of Flow
A company makes three products in four plants. Because of differences in labor
skills, different times are required to produce the products in the various plants.
The time requirements (minutes per unit) are shown in the table below. A dash
indicates that the product cannot be made at the plant. The demand for the
products is also shown. The total time available in each plant is 25 hours per week.
8

The hourly charge for labor is $10, 12, 9, and 13 for plants 1, 2, 3, and 4,
respectively. Not all the time available at a plant need be used. There is no labor
charge for time not used. Our problem is to construct and solve the network model
that minimizes the cost of manufacturing to meet the demand.

The example illustrates the use of gains to transform one type of flow to another.
The situation is like a transportation problem except the plants have capacities
measured in hours, and demand is in units. We use the gain factor transform the
flow as indicated in Fig. 21.

Using the gain factor to transform flow


The data for this situation is given in minutes per unit, so to compute the
appropriate gain factors, invert the quantities and multiply by 60 to find units per
hour. The gain for plant 1 producing product A is 2 units per hour. The complete
network model and the solution are shown in Fig. 22 and 23.

Model

Solution
Restrictions and Costs on Nodes
Consider again the distribution situation that began this chapter. The manager
learns that there is a problem with transshipments through Atlanta. The airport has
instituted a fee for each unit transferred and limits transfers to 100 units per week.
In addition, there is a spoilage of 10% of the units transferred.
The model of Fig. 2 does not allow this information to be included, because there
is no way to identify on a single arc the flow that represents the amount
transshipped. Restrictions can be placed on an arc, but not on the material passing
10

through a node. We accomplish this modeling task by dividing the Atlanta node
into two nodes as in Fig. 24 and adding an arc between. The flow that actually
passes through Atlanta now flows on arc 17, and any information related to that
flow is described by the parameters of the arc.

Including Node Information


Convex Costs and Concave Revenues
Consider an arc that carries flow between two nodes. We learn that the first 10
units of flow have a unit cost of $5, the next 5 units of flow have a cost of $8, and
any additional amount has a cost of $10.
With unit cost is increasing with flow, the arc cost is a convex function of flow.
Fig. 25 shows a linear model that has an arc for each level of unit cost. In cases
where the cost function is continuous, the model uses a piecewise linear
approximation with the number of arcs determining the accuracy of the
approximation. Concave cost functions cannot be addressed with linear models.
Nonlinear concave revenue functions can be handled in a similar way. The
revenues are represented with negative arc costs.

11

Representation of nonlinear costs


Undirected Edges
A commander has a flow problem involving points on the battlefield. The points
are numbered 1 though m, and the distance is given for every pair of points dij.
The distances are symmetric and unit cost is proportional to distance.
A problem of this nature might be represented as in Fig. 26 that shows the relations
between nodes as undirected edges. Since the flow model requires directed arcs,
the transformation in Fig. 27, shows an equivalent model with each edge replaced
by an oppositely directed pair of edges.

Model with nodes and edges


12

Model with directed arcs


Elimination of Arc Lower Bounds
Equivalent models with 0 lower bounds can replace models with lower bounds on
flow. Fig. 28 shows a general arc with the parameters lower bound, upper bound,
cost and gain. An equivalent representation of the arc with zero lower bounds is
also shown. The transformation requires modification of the external flows at both
ends of the arc. When arcs have equal upper and lower bounds, the effected arc
may be eliminated because its upper bound on flow will be zero.

Elimination of Lower Bounds

13

The term network flow program describes a type of model that is a special case of
the more general linear program. The class of network flow programs includes
such problems as the transportation problem, the assignment problem, the shortest
path problem, the maximum flow problem, the pure minimum cost flow problem,
and the generalized minimum cost flow problem. It is an important class because
many aspects of actual situations are readily recognized as networks and the
representation of the model is much more compact than the general linear program.
When a situation can be entirely modeled as a network, very efficient algorithms
exist for the solution of the optimization problem, many times more efficient than
linear programming in the utilization of computer time and space resources.

14

Potrebbero piacerti anche