Part 26: Picard’s Method for UPSC ISS: The Integration Loop Explained

Welcome to Part 26 and the opening of Module 7, updated for the UPSC ISS 2026 to 2027 cycle. Interpolation, numerical calculus, and difference equations are behind you. We now arrive at the pinnacle of the Numerical Analysis unit: Numerical Solutions of Ordinary Differential Equations. The goal changes fundamentally: given only a derivative $\dfrac{dy}{dx}$​ and one starting point, we reconstruct the curve itself. The opening tool is the elegant, integration driven Picard’s Method of Successive Approximations.

The method converts the initial value problem into the integral equation $y = y_0 + \int_{x_0}^{x} f(x, y)\,dx$ and iterates: replace the unknown $y$ inside the integral with the best approximation so far, integrate, and repeat until two successive approximations agree.

The exam context: 80 questions worth 200 marks in 2 hours, 2.5 marks per correct answer, about 0.83 deducted per wrong one, close to 20 questions from Numerical Analysis. Module 7 is the second largest module after Module 5 and the error trap zone of the paper.

This post is Part 26 of the Differential Equations Guide (Module 7), inside our UPSC ISS Numerical Analysis Complete Guide. New to the exam? Start at the UPSC ISS hub.

The Story of Dr. Meera and the Epidemic Curve

Dr. Meera, an ISS officer at the National Centre for Disease Control, studies an outbreak where the infection rate depends on both the day and the current patient count: $\dfrac{dy}{dx} = x – y$, with exactly 1 patient on Day 0, so $y(0) = 1$.

Her director needs the patient count on Day 0.2 to dispatch medical kits. Standard integration is blocked: $y$ appears inside its own derivative, and Meera does not know $y$ as a function of $x$. Her escape: assume a dummy value for $y$, integrate to get a better function, plug the better function back in, and loop. Two passes later she holds an accurate polynomial and reads off the answer, no calculator involved.

The Core Concept: The Integration Loop

For the first order equation $\dfrac{dy}{dx} = f(x, y)$ with $y(x_0) = y_0$, integrating both sides from $x_0$​ to $x$ gives the integral equation:$$y = y_0 + \int_{x_0}^{x} f(x, y)\,dx$$

The unknown $y$ still lurks inside the integral, so we iterate.

First approximation. Replace $y$ inside the integral by the known constant $y_0$​:$$y^{(1)}(x) = y_0 + \int_{x_0}^{x} f(t, y_0)\,dt$$

Second approximation. Feed $y^{(1)}$ back in:$$y^{(2)}(x) = y_0 + \int_{x_0}^{x} f\left(t, y^{(1)}(t)\right)dt$$

General step:$$y^{(n+1)}(x) = y_0 + \int_{x_0}^{x} f\left(t, y^{(n)}(t)\right)dt$$

The stopping rule. Halt when two successive approximations agree to the desired level of accuracy.

The Major Drawback: The UPSC Theory Trap

Picard’s method rewards you with a continuous polynomial equation rather than a table of discrete numbers, a genuine luxury. Its weakness is equally famous: the method is practically unsatisfactory whenever the required analytical integrations become severely difficult. If $f(x, y)$ carries terms like $e^{x^2}$ or tangled trigonometry, the integral cannot be evaluated by hand and the whole method stalls. This limitation is a direct theory question.

StatChakravyuh Pro Tips

1. The manual computation advantage. With calculators banned, Picard’s method is a friend, not a foe. Integrating simple polynomials like $t$ and $t^2$ mentally is faster and safer than the heavy decimal multiplications of the Runge Kutta family.
2. The dummy variable discipline. Always switch the integration variable to $t$, integrating $f(t, y)\,dt$ from $x_0$​ to $x$, so the limits never collide with the variables inside.
3. The rising degree pattern. Each pass of the loop raises the polynomial degree. If the first approximation tops out at $x^2$, expect the second to reach $x^3$. A mismatch between iteration count and degree is a quick sanity check.

Solved PYQ Masterclass

PYQ pattern (UPSC ISS 2017 cycle, as recorded in our PYQ archive): For the initial value problem $\dfrac{dy}{dx} = x – y$ with $y(0) = 1$, use Picard’s method to compute $y(0.2)$ correct to two decimal places.

Step 1: Parameters. $f(x, y) = x – y$, $x_0 = 0$, $y_0 = 1$.

Step 2: First approximation.$$y^{(1)}(x) = 1 + \int_0^x (t – 1)\,dt = 1 + \frac{x^2}{2} – x$$

Step 3: Second approximation. Substitute $y^{(1)}$ into the integrand: $t – \left(1 + \tfrac{t^2}{2} – t\right) = 2t – 1 – \tfrac{t^2}{2}$​. Integrate:$$y^{(2)}(x) = 1 + x^2 – x – \frac{x^3}{6}$$

Step 4: Evaluate at x=0.$$y(0.2) \approx 1 + 0.04 – 0.2 – \frac{0.008}{6} = 0.84 – 0.00133 \approx 0.8387$$

Final Answer: approximately 0.84, reached through two clean integrations and zero difference tables.

Common Traps to Avoid

Trap 1: Forgetting the leading $y_0$​ outside the integral. Every approximation begins with the constant $y_0$​ plus the integral.

Trap 2: Integrating with $x$ as both a limit and the integration variable. Use the dummy variable $t$ religiously.

Trap 3: A sign slip while simplifying the integrand of the second pass. Simplify fully before integrating, as we did with $2t – 1 – \tfrac{t^2}{2}$​.

Frequently Asked Questions

Take the Next Step with StatChakravyuh

Join the Free UPSC ISS WhatsApp Community. Every part of this series, plus the Module 7 Practice PDF with 60 MCQs, is delivered there first. 👉 Join the WhatsApp Community now

Enroll in the Free UPSC ISS Course. Structured lessons and practice sets built on the Practice Improve Repeat system, at zero cost. 👉 Enroll in the Free Course

Coming up in Part 27: when integration is impossible, we march instead. Euler’s Method, the simplest step by step solver, and its heavily tested error bounds.