Optimal staging — equal mass ratios (for same Isp): Foundations

Everything on the parent page rests on a small pile of symbols. If any one of them is fuzzy, the Lagrange proof collapses into meaningless squiggles. So we define each one from zero, in the order they depend on each other, and glue each to a picture.

1. Mass — the thing we are pushing

Before anything else: mass is just "how much stuff there is." We measure it in kilograms (kg). A full fuel tank has more mass than an empty one; a rocket with its payload has more mass than the payload alone.

We will constantly split a rocket's mass into three kinds of stuff:

Look at the figure: one tall bar, sliced into these three pieces.

Figure — Optimal staging — equal mass ratios (for same Isp)

Why do we need all three names? Because the whole optimisation is about a tension between them: fuel gives you speed, but fuel needs tanks (structure) to hold it, and those tanks are dead weight. Payload is the prize you are protecting. You cannot even state the problem without separating these three.

2. Exhaust speed $v_{e}$ and specific impulse $I_{s p}$ — how hard the engine pushes

A rocket moves by throwing gas backward. The faster it throws that gas, the more speed each kilogram of fuel buys you.

You will also meet specific impulse $I_{s p}$ . It is just $v_{e}$ rewritten in seconds: $v_{e} = I_{s p} g_{0}, g_{0} = 9.81 m/s^{2} .$

The parent needs "same $I_{s p}$ " as a hypothesis: if the engines differed, balancing the stages would be a different (harder) problem — see Variable specific impulse staging.

3. The logarithm $ln$ — why speed grows slowly as you add fuel

This is the symbol most readers fear, so we build it fully.

Start with a question: if throwing out the first kg of fuel gives some speed boost, does the hundredth kg give the same boost? No — and here is why. Each kg of fuel has to push not just the rocket, but all the fuel still on board. Early fuel pushes a heavy rocket (small boost); late fuel pushes a light one (big boost). The total gain is not a straight line — it curves and flattens.

The function that captures exactly this "flattening" is the natural logarithm, written $ln$ .

Look at the curve: it shoots up early, then bends flat.

Why this tool and not, say, a square root or a straight line? Because the rocket equation (next section) literally produces a logarithm from the physics of throwing mass backward. We are not choosing $ln$ for convenience — the physics hands it to us.

4. Mass ratio $R$ — the "leverage" of a burn

Now the central character.

$R = 1$ means "no fuel burnt, no change." $R = 4$ means the rocket started four times as heavy as it ended. Look at the two stacked bars — same rocket, before and after a burn.

Here $Δ v$ (read "delta-vee") means change in speed. The Greek $Δ$ ("delta") is standard shorthand for "the change in." $Δ v$ is what we are trying to maximise.

5. Two useful re-expressions of a single burn

The parent slices the burn two ways. Both come straight from the definition of $R$ .

Check the extremes: $R = 1$ gives $m_{f} = 0$ (no burn), and huge $R$ gives $m_{f} \to M$ (burned almost everything). Both sensible.

6. Structural fraction $ϵ$ — the "tax" on every tank

More on this in Structural fraction. From this definition you can rearrange to get structure from fuel: $m_{s} = \frac{ϵ}{1 - ϵ} m_{f},$ which is the exact expression the parent's proof uses to convert "fuel I chose to burn" into "structure I am forced to carry."

7. Payload fraction and the master constraint

First, the number we are scoring against.

Now the parent's central rule. Let $n$ be the number of stages (a plain counting number: $n = 2$ for a two-stage rocket, $n = 3$ for three, and so on). Label the stages $i = 1, 2, \dots, n$ from bottom to top, and let $R_{i}$ be the mass ratio of stage $i$ .

Why this exact product? Let us build one factor, then chain them — this is the sketch the parent skips.

Follow the mass as it flows down the rocket, one stage at a time. Write $M_{i - 1}$ for the mass entering stage $i$ (so $M_{0}$ is what the whole rocket starts with, and after the last stage burns and is dropped we are left with just $M_{n} = m_{p}$ ).

Stage $i$ burns fuel $m_{f i} = M_{i - 1} (1 - \frac{1}{R _{i}})$ — straight from §5, with $M = M_{i - 1}$ .
That fuel forces along structure $m_{s i} = \frac{ϵ}{1 - ϵ} m_{f i}$ — straight from §6.
After the burn and dropping the empty structure, the mass handed to the next stage is $M_{i} = M_{i - 1} - m_{f i} - m_{s i} = M_{i - 1} \frac{1 - ϵ R _{i}}{( 1 - ϵ ) R _{i}} .$

So each stage multiplies the mass by the factor $\frac{1 - ϵ R _{i}}{( 1 - ϵ ) R _{i}}$ . Do this $n$ times, from $M_{0}$ down to $m_{p}$ , and every factor multiplies onto the last — which is exactly the product $(⋆)$ . That is why it is a product and not a sum: each stage acts on whatever the stage below it left behind.

8. Lagrange multipliers — the tool for "maximise subject to a rule"

The parent asks: maximise total $Δ v$ while obeying the fixed payload fraction $(⋆)$ . That is a constrained optimisation — you can't just chase the biggest number, you must stay on the constraint surface.

Two ingredients get names in the method:

The symbol $\nabla$ ("nabla" or "gradient") is shorthand for "the list of slopes of a quantity in every direction you could nudge the $R_{i}$ ." Think of it as an arrow pointing in the direction that increases the quantity fastest.

Intuition Why this specific tool, and what

$λ$ means Ordinary "set the slope to zero" finds peaks of a free landscape. But here you are locked to a fence (the constraint). At the best allowed point, the direction you'd want to climb ( $\nabla f$ ) must point straight into the fence ( $\nabla g$ ) — otherwise you could slide along the fence and improve. Lagrange writes this as $\nabla f = λ \nabla g$ : the two arrows are parallel, and the number $λ$ ("lambda") is just how many times longer one arrow is than the other. Full method: Lagrange multipliers.

$λ$ also carries meaning: it is the exchange rate — how much extra $Δ v$ you would gain if the constraint were loosened by a tiny bit (e.g. if you were allowed a slightly smaller payload). A big $λ$ means the constraint is "expensive"; the payload is really holding you back.

You do not need to master the machinery to trust the parent's punchline: because every stage sees the same $ϵ$ , each stage's optimality equation is identical, so all the $R_{i}$ must come out equal. That is the whole theorem, and now you know what every symbol in its statement means.

Prerequisite map

Read it top to bottom: raw mass ideas feed $R$ and $ϵ$ ; $R$ plus $ln$ plus $v_{e}$ give the rocket equation; the constraint plus Lagrange give the theorem.

Equipment checklist

Cover the right side and test yourself. If any answer surprises you, re-read that section before the parent page.

What does the natural log

ln x

ask?

"To what power must

e \approx 2.718

be raised to give

x

?" — and it grows ever more slowly.

Why must the mass ratio

R

always be

\geq 1

Burning fuel only removes mass, so the "after" mass is never bigger than the "before" — the ratio can't drop below 1.

State the Tsiolkovsky rocket equation for one burn.

Δ v = v_{e} ln R

What is the difference between

v_{e}

and

I_{s p}

Same physical fact;

v_{e} = I_{s p} g_{0}

v_{e}

is a speed (m/s),

I_{s p}

is it rewritten in seconds.

Define structural fraction

ϵ

in words.

The share of a stage's structure-plus-fuel mass that is dead structure:

ϵ = m_{s} / (m_{s} + m_{f})

Why does

ϵ > 0

make staging matter?

With

ϵ = 0

every fuel split gives the same speed; only the structural "tax" makes the optimal split genuinely better than others.

What does

\prod

mean, versus

\sum

\prod

= multiply all terms together;

\sum

= add all terms together.

Why is the master constraint a product and not a sum?

Each stage multiplies the incoming mass by its surviving-fraction factor; chaining

n

stages multiplies

n

factors together.

Fuel burned by a stage of start-mass

M

and ratio

R

m_{f} = M (1 - 1/ R)

In Lagrange's method, what are

f

and

g

f

is the objective you maximise (

\sum ln R_{i}

);

g

is the constraint you must hold fixed (

\sum ln

of the per-stage factors

= ln (m_{p} / M_{0})

What does the multiplier

λ

measure?

The exchange rate — how much extra

Δ v

you'd gain if the constraint were loosened slightly.

Why use Lagrange multipliers here rather than plain calculus?

We must maximise

Δ v

while staying on a fixed constraint surface; Lagrange enforces "the gradients

\nabla f

and

\nabla g

line up" at the best allowed point.

Why does equal

ϵ

across stages force equal

R_{i}

Each stage's optimality equation depends only on its own

R_{i}

and the shared

ϵ

, so all stages solve the identical equation — hence equal ratios.