Skip to main content

Load Balancing

Maximum Flow · Min-Cut · Polynomial

Transmission congestion costs North American grids $5–$12 billion annually in inefficient generation redispatch and lost renewable curtailment. Every 15 minutes, a grid operator must route power flows from generators to load centers through a capacitated transmission network. This is the Maximum Flow problem—solvable in polynomial time by Edmonds-Karp in O(VE²)—where transmission line ratings and transformer limits determine the bottleneck capacity.

Where This Decision Fits

Energy operations chain — the highlighted step is what this page optimizes

Plant Siting Facility location
Transmission Network Network design
Unit Commitment Generator scheduling
Load Balancing Distribution
Market Trading Price optimization

The Problem

From power grids to graph theory

A regional power grid connects 6 generators (sources) and 8 demand districts (sinks) through 15 transmission lines, each with a MW capacity limit set by thermal ratings and transformer constraints. The objective is to find the maximum power flow from all generators to all demand centers and identify the bottleneck (minimum cut) that limits total throughput.

This maps directly to the Maximum Flow problem on a directed graph. By the Max-Flow Min-Cut Theorem (Ford & Fulkerson, 1956), the maximum achievable flow equals the capacity of the minimum s-t cut—the smallest set of lines whose removal disconnects supply from demand.

Energy Domain Max-Flow Model
Generator Source node
Demand district Sink node
Transmission line Edge with capacity c(u,v)
Thermal rating (MW) Edge capacity
Power flow Flow f(u,v)
Congested corridor Min-cut edge
Max-Flow  —  Polynomial  (Edmonds-Karp: O(VE²))

Mathematical Formulation

The Maximum Flow LP relaxation

Objective maximize   |f| = Σ(s,v)∈E f(s,v)   // total flow from super-source

Subject To capacity:   0 ≤ f(u,v)c(u,v)   ∀ (u,v) ∈ E // flow cannot exceed line rating
conservation:   Σ(u,v)∈E f(u,v) = Σ(v,w)∈E f(v,w)   ∀ v ≠ s,t // Kirchhoff's current law
skew symmetry:   f(u,v) = −f(v,u)   ∀ (u,v) ∈ E

Interactive Solver

Choose a scenario, pick an algorithm, and watch the flow

Network Flow Solver

All 15 transmission lines operational. Normal demand pattern across 8 districts.

<50% utilized
50–85% utilized
>85% utilized
Min-cut edge
Algorithm Max Flow (MW) Augmenting Paths Min-Cut Edges Time
Press Solve to run all algorithms
Ready — select a scenario and click Solve

Solution Algorithms

Three approaches to finding maximum flow

Augmenting Path

Edmonds-Karp

O(VE²)

The Edmonds-Karp algorithm is a specific implementation of the Ford-Fulkerson method that uses breadth-first search (BFS) to find augmenting paths. By always selecting the shortest augmenting path, it guarantees a polynomial bound of O(VE²), making it the standard choice for moderate-size networks.

In the energy context, each BFS traversal finds a transmission path from generators to loads that still has available capacity, pushing as much power through as the bottleneck link allows.

Augmenting Path

Ford-Fulkerson

O(E × |f*|)

The original augmenting-path method uses depth-first search (DFS) to find paths from source to sink with residual capacity. While simple, its worst-case complexity depends on the max-flow value |f*|, which can make it non-polynomial for irrational capacities.

For integer-capacity transmission networks (MW ratings are always integers), Ford-Fulkerson terminates in at most |f*| iterations, each augmenting flow by at least 1 MW.

Preflow

Push-Relabel

O(V²E)

Unlike augmenting-path methods, Push-Relabel maintains a preflow (flow that may exceed conservation at intermediate nodes) and uses local push and relabel operations to achieve feasibility. The highest-label variant achieves O(V²√E) in practice.

For dense transmission networks with many parallel paths, Push-Relabel often outperforms BFS-based methods due to its local operation model and better cache behavior.

Complexity Comparison

  • Edmonds-Karp: O(VE²) — polynomial, BFS guarantees shortest augmenting paths
  • Ford-Fulkerson: O(E × |f*|) — pseudo-polynomial, DFS may explore long paths
  • Push-Relabel: O(V²E) — polynomial, best for dense graphs with high connectivity
  • Max-Flow Min-Cut Theorem: the maximum flow from s to t equals the minimum cut capacity separating s and t (Ford & Fulkerson, 1956)
  • Practical note: real grid solvers use DC power flow approximations and PTDF matrices rather than pure graph algorithms, but the combinatorial structure is identical

References

Key literature on network flow and power systems

Ford, L.R. & Fulkerson, D.R. (1956).
"Maximal flow through a network."
Canadian Journal of Mathematics, 8, 399–404.
Edmonds, J. & Karp, R.M. (1972).
"Theoretical improvements in algorithmic efficiency for network flow problems."
Journal of the ACM, 19(2), 248–264.
Goldberg, A.V. & Tarjan, R.E. (1988).
"A new approach to the maximum-flow problem."
Journal of the ACM, 35(4), 921–940.
Hobbs, B.F., Rothkopf, M.H., O'Neill, R.P., & Chao, H. (Eds.). (2001).
"The Next Generation of Electric Power Unit Commitment Models."
Springer, International Series in Operations Research & Management Science.

Need to optimize energy grid
transmission flows?

Get in Touch
Data shown is illustrative. This is a simplified model for educational purposes.
 Home
ESC