Temporarily taking over a small office, Hua Luogeng accepted writing paper and a pen from another young staff member.
The room was small, containing only a wooden desk, a red telephone, a row of filing cabinets, and a brazier for warmth.
Hua Luogeng pulled out a chair and sat down, spreading the writing paper flat on the desk.
"Director Hua, I'm terribly sorry for the trouble."
Researcher Zhou brought over the relevant research data and spoke with an apologetic look on his face.
"It's no trouble at all."
Hua Luogeng shook his head, took the research data from him, and set it on the desk. "I can't guarantee I'll be able to solve this problem quickly. All I can do is give it a try first."
Researcher Zhou said, "Thank you for your hard work. If you need anything, just call and tell me."
"All right."
Hua Luogeng nodded in acknowledgment, then unfolded the experimental data and began looking through it.
He didn't start writing immediately. Instead, he went through the report again from beginning to end.
The iterative algorithm for predicting the Hongqi-1 missile's trajectory essentially expanded the equations of relative motion between the missile and its target into a series according to their parameters, then used successive iterations to approximate the true solution.
The framework of the algorithm had been extracted from Big Brother's S-75 materials.
But they had held something back in the crucial control law section, providing only the formulas without their derivations, much less an analysis of convergence.
As a result, when the Hongqi-1's interception target maneuvered at high speed and altitude, the error function sequence for its trajectory deviation developed a divergent component from the fourth step onward that conventional relaxation iteration could not suppress.
In other words, guidance and hit accuracy would drop sharply, leaving it vulnerable to interference.
"The function sequence itself behaves abnormally at the convergence boundary. We need to find a way to account for it."
"Would changing the basis work?"
Looking at the experimental data on the desk, Hua Luogeng pushed up his glasses and thought it over carefully as he began working out calculations on a blank sheet of paper.
He quickly ruled out one promising avenue.
The equations on the paper ultimately still pointed to divergence.
"Relaxation iteration doesn't work, and changing the basis treats the symptoms, not the root cause. This doesn't look like a problem with computational precision. It looks more like an inherent problem with the structure of the algorithm itself."
Muttering to himself, Hua Luogeng leaned back in his chair, his fingers absently rubbing the warm rim of his enamel teacup.
Relaxation iteration, Chebyshev acceleration, Aitken extrapolation... he had checked all these methods repeatedly, and every one of them treated the symptoms rather than the root cause.
That was because divergence couldn't be plugged with numerical tricks; the algorithm's structure itself had an inherent flaw.
He needed to get around the threshold of reflexivity and construct a control sequence in a non-reflexive space that could suppress the divergence.
But he had been unable to find a breakthrough on that step.
Although he knew the problem might not be so easy to solve, Hua Luogeng was still growing anxious.
After all, the motherland's territory and national strategic bases were under others' surveillance at all times, leaving no room for delay.
He rubbed the slightly sore bridge of his nose, took off his glasses, and blinked to ease his tired eyes.
Perhaps the fatigue from his nonstop journey had finally overcome his will, or perhaps the warmth of the charcoal fire in the room made it easy to relax.
His consciousness seemed to be gently cradled by an exceedingly gentle hand. His eyelids drooped heavily, and his breathing gradually grew slow and deep.
In a daze, Hua Luogeng felt as though he had dreamed.
In the dream, there were no red-brick, tile-roofed buildings, no storeroom, only a yellowed book floating in the void.
The book's cover was indistinct, but the writing on its title page was perfectly clear.
"Cleverness lies in learning; genius lies in accumulation."
Hua Luogeng froze. It was his writing, from the preface he had written last year while compiling teaching materials for USTC's students.
Before he could react, the writing rose from the page, turning into a soft, lustrous trace of ink that slowly unfurled in the void.
[The function sequence {fₙ} is defined on E. There exists a nonnegative function sequence {φₙ} that converges uniformly on E, such that |fₙ(x)|≤φₙ(x) holds for ∀n∈ℕ, ∀x∈E; then {fₙ} converges uniformly on E...]
The hazy writing appeared in his dream, and Hua Luogeng's eyes widened.
He tried to reach out and touch the writing, but found he couldn't reach it no matter what he did. He tried to call out, but couldn't make a sound.
But he could "read" and "see."
He saw the idea of the Frenet frame reduced to a norm gradient and transferred into function space.
He saw how the weak compactness of the Banach-Alaoglu theorem tightened on the boundary layer.
He saw how the dual action δ_ij ensured orthogonality in every direction, locking the error components firmly onto their respective coordinate axes so they did not interfere with one another...
The formulas flowed into his mind like running water, as though reaching a deeper resonance with his brain across the long years.
He didn't know how much time had passed before he jolted awake at the wooden desk.
He raised his head. Nothing around him had changed: the writing paper was still on the desk, and the coals in the brazier flickered with a dark red glow, giving off scattered points of warmth.
"The control sequence... the control sequence..."
Hua Luogeng muttered to himself, recalling the scene that had seemed like a dream. Suddenly, he grabbed his ballpoint pen and quickly wrote line after line on the paper.
"Error control function"
"In the iterative calculation, construct a sequence of explicitly computable 'control functions' φₙ(x), such that they decay monotonically at the near-boundary intersection δ(x) and bound the original iterative error term |fₙ(x)-f(x)| everywhere..."
"...If φₙ itself converges uniformly to zero on the near boundary, then the original iterative function sequence converges uniformly over the entire domain."
The pen scratched across the paper.
He wrote faster and faster, the light beneath his glasses growing brighter and brighter.
The divergent components that relaxation iteration had repeatedly suppressed, only for them to rebound again and again, were now broken down into several independent parts within the control sequence's framework, located one by one and quantified one by one.
"Take the near-boundary distance function δ(x) as the relative distance between the missile and the target, and construct the control function φₙ(x)=C(-n·δ(x))."
"At the near-boundary intersection, δ(x)→0, and the control function remains bounded; in the far field, δ(x) increases, and the control function decays exponentially—the entire domain is under control!"
Two hours later, the pen in Hua Luogeng's hand came to a stop.
He stared quietly at the pages in his hand, experiencing a feeling he couldn't quite describe.
In a daze, had he just seen the answer to this problem?
But now wasn't the time to think about that. He picked up the pages from the desk, pulled himself together, and checked them carefully twice. After confirming that this method could solve the Hongqi-1 missile's divergence problem, he picked up the hand-cranked telephone on the desk and quickly made a call.
Before long, hurried footsteps sounded outside the room.
Researcher Zhou pushed open the door and came in with two researchers.
"Director Hua."
"Take a look and see if it meets your requirements." Hua Luogeng stood up and handed over the pages he had put in order.
"All right!"
Trembling slightly, Researcher Zhou took the pages and read through them line by line.
"...Take the near-boundary control function φₙ(x)=C(-n·δ(x)), where δ(x) is the relative distance between the missile and target. The iterative error can then be strictly controlled within 10⁻³, meeting the guidance accuracy requirement."
"This is it!"
"Director Hua, you're truly amazing! You actually came up with such an effective method!"
Researcher Zhou slapped the desk, his weary face instantly alight with excitement and exhilaration.
"Check the calculations first."
Hua Luogeng spoke calmly. "Use the number seven near-boundary test case you gave me. Have three people calculate it simultaneously on hand-cranked calculators and cross-check the results."
"Replace the original constant majorant series Mₙ with this near-boundary control function φₙ(x)=C(-n·δ(x)), and run another three iterations."
"All right! I'll go right now!"
Three hours.
For three whole hours, the storeroom held only the clicking of hand-cranked calculators and the scratching of pen tips across paper.
Three calculators, six hands. New blood blisters formed from turning the cranks, but no one cried out in pain, and no one stopped.
Researcher Zhou stared unblinkingly at the final verification results.
First iteration, error sequence: 0.042, 0.031, 0.028, decreasing.
Second iteration, error sequence: 0.028, 0.019, 0.015... decreasing.
Third iteration, error sequence: 0.015, 0.009, 0.006... decreasing.
The fourth step, the fifth, the sixth...
Convergence!
Stable convergence!
No divergence! No oscillation! No exponentially growing error!
The interference that had trapped them at the near-boundary intersection for months had been held firmly in check by that seemingly simple exponential function.
It was as though a fierce tiger had been locked in a cage.
"Director Hua!"
Seeing this, Researcher Zhou raised his head. His voice choked and his eyes reddened; he could no longer hold himself back. "The error at the seventh step has already been reduced below 10⁻³. It meets the guidance accuracy requirement."
"Our Hongqi-1... its guidance commands can be finalized!"
"In less than half a year, we'll have air defense missiles of our own!"
The room fell silent for an instant, then erupted in a burst of cheers held back to the utmost.
The young calculators whose eyes were red from sleepless work and the engineers whose fingers had been rubbed raw embraced one another and patted one another on the back.
Everyone was celebrating, their voices trembling violently with excitement, their hearts pounding in their chests, almost hard enough to shatter their ribs.
Some laughed, some cried, and some buried their faces in the calculation sheets produced by the hand-cranked machines, silently shedding tears.
Hua Luogeng stood in the calculator room with a smile on his face.
Although he still didn't know what that dream had been about, the finalization of the Hongqi-1 filled him with excitement.
Half a year later, a Hongqi-1 surface-to-air missile, first deployed with the 1st Surface-to-Air Missile Battalion, precisely shot down a U-2 high-altitude reconnaissance aircraft in the night sky over the Baotou area. The pilot was captured alive, and all the electronic equipment aboard the enemy aircraft was seized.
When the news broke, the international world was in an uproar.
The Pentagon convened an emergency meeting overnight. They couldn't understand it!
The U-2 flew at an altitude of twenty thousand meters. Before this, the other side hadn't had a single air defense missile that could reach that height.
Kosmo's experts were equally silent. When they withdrew, they had asserted that Hua Country wouldn't be able to produce a decent air defense missile in ten years.
And now, from the project's launch to shooting down the U-2, the other side had taken less than two years.
P.S. This is the second update. Please give me monthly votes, recommendation votes, continued readership, and comments. Newbie Cat wants everything! Thank you~