Network Models

A network model represents a system as a set of nodes (vertices) connected by arcs (edges), where each arc carries a flow, cost, time, or capacity. Networks are used to model transportation routes, communication links, pipeline systems, project activities, and supply chains. A directed arc (i, j) has a direction of flow, while an undirected edge can be traversed in either direction. Each arc typically has a capacity u_ij (the maximum flow it can carry), a cost c_ij per unit of flow, and a traversal time t_ij. The most common network problems are: the shortest path problem, which finds the minimum-cost or minimum-time route from a source to a destination; the maximum flow problem, which finds the largest flow that can be sent from a source s to a sink t without exceeding arc capacities; the minimum-cost flow problem, which sends a required flow at minimum total cost; the transportation and assignment problems, which distribute goods from sources to destinations; and the critical path method (CPM) and PERT, which schedule project activities represented as a network of precedence relationships. The shortest path problem is solved by Dijkstra's algorithm when all arc costs are non-negative, or by Bellman-Ford when negative costs are present. Maximum flow is solved by the Ford-Fulkerson method and its variants (Edmonds-Karp, Dinic). The max-flow min-cut theorem states that the maximum flow from s to t equals the capacity of the minimum s-t cut, which separates the network into two parts with the smallest total capacity. In project networks, the critical path is the longest path through the network; any delay on a critical activity delays the entire project, while non-critical activities have float or slack. Network models assume that arc parameters are known and deterministic; when travel times or demands are uncertain, stochastic or robust versions are used. The main limitations are the assumption of linear, additive costs and the difficulty of representing complex real-world constraints such as vehicle routing with time windows.

Governing FormulaShortest path (Dijkstra): d(v) = min over edges (u,v) of [ d(u) + w(u,v) ] Maximum flow conservation: Σ_in f(i,j) - Σ_out f(j,k) = 0 for intermediate nodes Capacity constraint: 0 ≤ f(i,j) ≤ u(i,j) Max-flow min-cut theorem: max flow = min cut capacity Minimum-cost flow objective: minimize Σ c(i,j) · f(i,j) Subject to: Σ f(i,j) - Σ f(j,k) = b(j) (node balance) Flow cost: Z = Σ c(i,j) · f(i,j) Project network: earliest start ES(j) = max over predecessors [ ES(i) + t(i,j) ] Total float: TF(i) = LS(i) - ES(i) = LF(i) - EF(i) Critical path: all activities with TF = 0

Knowledge Check

10 Questions

1.What is a node in a network model?

2.What is the difference between a directed arc and an undirected edge in a network?

3.What does the capacity u(i,j) of an arc represent in a maximum flow problem?

4.Which algorithm is used to find the shortest path from a source to all other nodes when all arc costs are non-negative?

5.What does the max-flow min-cut theorem state?

6.In a project network, what is the critical path?

7.An activity has an early start of day 5, an early finish of day 9, a late start of day 7, and a late finish of day 11. What is its total float?

8.In the minimum-cost flow problem, what is the objective?

9.Which assumption underlies the standard deterministic network models?

10.A transportation network has a single source with supply 50 and a single sink with demand 50. The maximum flow found is 40. What does this indicate?