A staggered system: start from the equation with one unknown
One equation has one unknown, another two, the third three; the question asks for one number or a combination such as .Trap: giving a single value when a combination is asked.
In the textbook: Ch. 6.2 Das Lösen linearer Gleichungssysteme
Two or three unknowns and as many equations: the system is solved by removing the unknowns one at a time until the last is left, then substituting it back.
You need a system of equations when there are several unknowns and no single condition fixes them one by one. That happens whenever only totals are known: how much was sold in all and how much money is in the till, how much raw material went into the whole batch. In the exam such a system is either a task of its own or one step inside a bigger task.
A small café sold 10 drinks in one day, coffee at 3 € and tea at 2 €. In the evening the till holds 26 €.
How many of these drinks were coffee?
One condition is not enough: 10 drinks can be 5 + 5 and also 6 + 4. The takings alone decide nothing either. The answer comes only when you take both conditions at once.
If all 10 drinks were tea, the till would hold €. Each coffee instead of a tea adds 1 €, and the till holds 6 € more. So there were 6 coffees and 4 teas. Check: . With two unknowns this reasoning is handy. With three you need a safer method: that is what this article is about.
Check yourself. A task at exam level: find if , and .
Call the number of cups of coffee and the number of cups of tea . Each condition in the text becomes an equation:
Two equations in the same unknowns, taken together, are called a system of linear equations (German: lineares Gleichungssystem). “Linear” means that the unknowns are only multiplied by numbers and added: there is no and no . The equations are numbered with Roman numerals I, II, III in the order they are written.
Let us test two pairs of numbers:
A set of numbers, one for each unknown, that makes all the equations true at once is called a solution of the system.
Guessing was lucky here. With three unknowns and numbers like 1203 guessing does not work; you need a method that always leads to the answer.
If the second equation had no , the number could be read off at once. Let us remove , using only the fact that both equations are true. This is how Gaussian elimination (“Gauß-Verfahren”) works: the unknowns are removed one after another.
This step is written “II I”: subtract I, multiplied by 2, from II. Beside the equations it helps to keep a table of numbers only, because and stand in the same places in every row and only the numbers in front of them change. Such a table is called the augmented matrix (German: erweiterte Koeffizientenmatrix): a row is an equation, a column belongs to one unknown, the right-hand side stands behind the bar. Below, the same two café equations are written as such a table.
row I times 2
In front of there is a 2 in II and a 1 in I. For to disappear it needs the same number in front of it, so we multiply I by 2.
, that is .
Why ask whether you can go back? If a step can be undone, the system before it and after it have the same solutions: none is lost, none is added. In the café it goes like this: add to the new row again and you get II. Each of the three steps that never change the solutions can be undone in the same way:
Every step acts on both sides of the equation, the right-hand side included. Removing an unknown with such steps is called elimination (German: Elimination).
If the coefficients are not 1. In the café there was a 1 in front of in I, and multiplying one equation was enough. Usually you have to multiply both.
I , II . Remove .
In front of stand 2 and 3. A number divisible by both 2 and 3 is 6.
: .
: .
, the terms with : .
The terms with : .
On the right: . This gives , so .
In I: , that is .
, . Check with II: .
If the first equation has no at all, swap it with an equation that has (step 1).
Trap: subtracting on the left but not on the right. In the café this gives , then , and the check with II gives instead of 26. A negative number of cups already hints at the mistake.
Your turn1 of 5
From I: and II: , is removed with the step . Which equation results?
Elimination serves one purpose: to get an equation with only one unknown left. In the first exam task such an equation is already in the question:
Gegeben ist das lineare Gleichungssystem in :
Berechnen Sie .
The task asks for , so all three unknowns are needed. The task is recognised by the shape of the system: the bottom equation has one unknown, the middle one two, the top one three. Such a system is called staggered, or in row echelon form: each following equation starts one unknown further to the right. There is nothing to eliminate here, that work has been done. What remains is back substitution (German: Rückwärtseinsetzen): start with the equation that has only one unknown; the number found turns the neighbouring equation into an equation in one unknown, and so on. Here that equation is at the bottom, so you go from the bottom up.
2x₃ = −52
.
Divide both sides by 2: .
The system , , ; the sum is wanted. The equation with one unknown is at the top, so you start at the top.
, so .
, that is .
.
, that is .
, so .
.
From the bottom up. The system I , II , III : the equation with one unknown stands at the bottom. Find .
Your turn2 of 5
For : , , . What is ?
Your turn3 of 5
For : , , . What is ?
Traps. The wrong options of the task above come about like this:
Substituting from the bottom up only works when the steps are already there. In the second exam task there are no steps yet. You first build them by elimination, as in the café section:
Gegeben ist das lineare Gleichungssystem
Bestimmen Sie den Wert von .
Here three equations are given, each with all three unknowns, and the question is about one of them. There are no steps yet: they are made by elimination. The first equation removes from the second and third. Two equations in and remain, which is the task from the café section. From them is removed, and the last row contains only . The new row after a step gets a prime: II′ means “the new second equation”. A row changed a second time gets two primes: III″.
II − 2·I
In front of there is a 2 in II and a 1 in I, so we compute II I.
: .
The terms with : . The terms with : . On the right: .
Row II′: .
The check with III takes half a minute, and it is needed: and also came out along the way, and both are among the wrong options.
The system I , II , III . Find .
All the numbers in I are divisible by 2. Divide (step 2 with the number ): .
II I: is .
, , on the right . Row II′: .
III has no , so it is left alone.
In front of stand in II′ and 1 in III. : .
II′ III: , , on the right .
, .
From III: . From I: , . Check with II: .
Remove from III′. The system from “Check yourself”: I , II , III . II I gives II′ . III I gives III′ . Find .
Your turn4 of 5
Find if , and .
Your turn5 of 5
Find if , and .
Traps. The wrong options of the task above come about like this:
The topic in a minute
Trap: giving a single value when a combination is asked.
Trap: giving or when is asked; not recomputing the right-hand side.
On the exam formula sheet there is no Gaussian elimination: you have to remember the steps yourself.
The systems in this article each had exactly one solution. There can also be none, or infinitely many, which is the subject of 'How many solutions a system has: none, one or infinitely many'.
Was this article helpful?