All revision notes topics

Multiple sources and sinks, restricted capacities and optimal flowEdexcel A-Level Further Maths: Revision notes

Section 1

Multiple sources and sinks

A network may have several sources (for example, several depots) and several sinks (several shops). Add a super source SS with an arc to each source, and a super sink TT with an arc from each sink. Then solve as a normal single-source, single-sink problem. The capacity of each new arc is the supply of that source, or the demand of that sink. If a source can supply any amount, give its arc infinite capacity (or a number larger than any possible flow). Example: depots P (supply 9) and Q (supply 6): arcs SP of capacity 9 and SQ of capacity 6.

Key termssuper sourcesuper sinksupplydemand
Common mistake

Putting the demands of the sinks on the arcs from the super source. Supplies go on the super source arcs and demands on the super sink arcs.

Section 2

Vertices with restricted capacity

Sometimes a vertex CC can handle only a limited total flow, say 77. Split it into two vertices, CinC_{\text{in}} and CoutC_{\text{out}}, joined by an arc CinCoutC_{\text{in}}C_{\text{out}} of capacity 77. Every arc that ended at CC now ends at CinC_{\text{in}}, and every arc that started at CC now starts at CoutC_{\text{out}}. All flow through CC must use the new arc, so the throughput is at most 77. The restriction counts as an arc in cuts: for example, the cut {S,A,B,Cin}\{S,A,B,C_{\text{in}}\} includes the arc CinCoutC_{\text{in}}C_{\text{out}}. Do not simply reduce the capacity of one arc: if CC has several outgoing arcs, the others stay unrestricted.

Key termsrestricted capacitythroughput

Section 3

Lower and upper capacities

Some arcs must carry at least a minimum flow as well as at most a maximum. Write each arc as [lower,upper][\text{lower}, \text{upper}]. A flow is feasible if on every arc lower≤flow≤upper\text{lower}\le\text{flow}\le\text{upper} and flow in equals flow out at each intermediate vertex. Labelling with lower capacities: the arrow in the direction of the arc shows upper −- flow (how much it can increase); the arrow in the opposite direction shows flow −- lower (how much it can decrease). Start from a feasible flow and augment as before; the flow never drops below any lower capacity.

Key termslower capacityupper capacityfeasible flow
Common mistake

Using the whole flow as the backward label. With a lower capacity, the backward label is flow minus lower capacity.

Section 4

Cuts with lower capacities

For a cut with sets XX (containing S) and YY (containing T): capacity of cut=∑(upper capacities of arcs from X to Y)−∑(lower capacities of arcs from Y to X).\text{capacity of cut}=\sum(\text{upper capacities of arcs from }X\text{ to }Y)-\sum(\text{lower capacities of arcs from }Y\text{ to }X). The arcs from YY to XX matter now: their minimum flows must return across the cut, which uses up some of the flow that can cross forwards. With all lower capacities zero, this is the earlier rule. Example: arcs SA [1,11], BT [2,8] go out of X={S,B}X=\{S,B\}, and AB [3,9] comes into XX from YY: capacity 11+8−3=1611+8-3=16.

Key termscapacity of a cut
Exam tip

Check every arc in both directions when you work out a cut: forward arcs add the upper capacity; backward arcs subtract the lower capacity.

Section 5

Finding the optimal flow and improving a network

To find the optimal (maximum) flow with lower capacities: start from any feasible flow, augment with the labelling procedure until T cannot be labelled, then confirm with a cut whose capacity equals the flow value (max-flow min-cut). To improve the flow, change an arc in the minimum cut. Raising the upper capacity of an arc from XX to YY in the minimum cut, or lowering the lower capacity of an arc from YY to XX, increases the cut capacity. Changes to arcs elsewhere do nothing. The gain stops when another cut becomes the smallest, so compare with the next-smallest cut capacity.

Key termsoptimal flow
Exam tip

After an improvement, check the next-smallest cut. The maximum flow rises only until it meets that cut.

That's the notes covered.

Carry on to the next subtopic.

Exam questions on Multiple sources and sinks, restricted capacities and optimal flow

  1. Two depots, P and Q, supply goods to two shops, U and V, through junctions A and B. Depot P can supply at most 9 units and depot Q at most 6 units. Shop U can receive at most 7 units and shop V at most 8 units. The directed roads and their capacities, in units, are: PA 6, PB 4, QB 5, AU 4, AV 3, BU 2 and BV 6. To find the greatest total delivery, a super source S and a super sink T are added.
    A flow of value 1414 is found in which depot P supplies 99 units. Show that this flow is a maximum flow, and find how much of the supply at depot Q is not used.2 marks
  2. A pipeline network has source S, sink T and junctions A, B, C and D. The directed pipes and their capacities, in units per hour, are: SA 8, SB 7, AC 6, AT 3, BC 5, CT 6, CD 4 and DT 5. Junction C can handle at most 7 units per hour in total.
    The restriction at junction C is removed. Find the new maximum flow and justify your answer.2 marks
  3. A network has source S, sink T and vertices A, B and C. Each arc has a lower and an upper capacity, written [lower, upper]: SA [3, 9], SB [2, 6], AB [1, 4], AC [0, 5], BC [2, 6], BT [0, 4] and CT [4, 10].
    Find the capacity of the cut with X={S,B}X=\{S,B\} and Y={A,C,T}Y=\{A,C,T\}, and of the cut with X={S,A,B,C}X=\{S,A,B,C\} and Y={T}Y=\{T\}. What can you deduce about the maximum flow?3 marks
See the full worksheet

Written by the Exaim team, led by Shaun Daswani (Head of Upper Secondary, Improve ME Institute; MSc Financial Mathematics, Imperial College London; BSc, UCL) and Jason Daswani (operational lead, Improve ME Institute; LSE).