Iosif I. Perel’man

Iosif I. Perel’man


Thursday, August 28, 1924

Passed away: 

Thursday, October 6, 1994

Iosif Perel’man, in full Iosif Isaakovich Perel’man, was born in Moscow. He was drafted to the front after finishing a secondary school. Of low stature, frail, even stubby (throughout his life), and with short-sightedness, Perel’man caused bewilderment among recruits. To the question “How did you get here?” he learned to answer: “On foot.”

Perel’man did not like to talk about the Great Patriotic War, which he finished as a sergeant. He was jokingly proud of his wound in the heel: by Soviet troops! All his life, Perel’man had been limping a bit. During serious discussions, he called a war a survival game with special rules, where the stake was often life: one’s own or someone’s else. There was no coldness and detachment in his words; Perel’man knew it firsthand. He was like that in everything: a powerful rational intellect without any sentimentality and pathos.

After the War, Perel’man received a long medical treatment, graduated from Moscow Power Engineering Institute, and worked in the “box” (designed and tuned drive control systems for radar stations on warships). He finished postgraduate studies at the Institute of Automation and Remote Control (IARC), the USSR Academy of Sciences. In 1956, Perel’man joined Meerov’s Laboratory at IARC. In 1967, he headed an independent group, which merged Itskovich’s Laboratory in the late 1980s. According to those who worked together and knew him, Perel’man was very talented and quick to grasp the essence, not only in narrow professional areas. To many colleagues, Perel’man seemed to be extremely rational: he did not like to discuss unknown things without clear explanations. It was not easy to explain something to Perel’man! He liked to argue and was a true master of scientific debate. Perel’man never concealed his opinion about the interlocutor’s abilities, though making it clear without explicit offense. He had never humiliated or offended anyone personally. Perel’man had a high opinion of his abilities (however, with full right).

For 40 years of his scientific activities, as can be traced now, Perel’man developed one central topic: the theoretical and engineering justification of control algorithms using the object’s dynamic model. During those years, the corresponding area of control science was born and had been gradually formed with almost complete exhaustion. His research began in the 1950s with the principle of a self-adjusting program (apparently, the first idea of parametric recurrent identification). It continued with the phenomenological approach to identification in control problems (in the 1970s). Finally, the dynamic optimization algorithm was developed (in the 1980s). Currently, the last direction, independently pioneered by him, has grown in the West into the so-called Model Predictive Control (MPC). It has underlaid R&D works on improved industrial control, bringing the suppliers of commercial solutions multimillion profits. For the first time, this problem integrated the two-scale model principle, sliding interval prediction, and control loop simulation. Perel’man’s idea can be considered one of the best in modern control science.

Perel’man was distinguished by his engineering approach to problems (maybe, to life, too). He did not like beautiful formulations, especially without any matter behind them. Perel’man revealed them immediately regardless of personality (not always for his benefit).

Perel’man was principled and demanding to the quality of R&D works, whether his own or of his employees. It was impossible to negotiate intermediate results with him: only 100%-completed works were released. Perhaps, that explained a relatively small number of publications and implementations. Though, Perel’man managed to do a lot. The main practical and mature results were adopted in metallurgy, cement industry, and petrochemical industry.

Perel’man was generous with ideas. Many studies of others began in conversations with him, and they were not ashamed to admit it. He gathered such a “harvest” not too diligently. Perel’man was not hard to get along with, but he did not want “to make a fuss,” and his superiors did not like him. Perel’man became Doctor of Engineering and Professor after a long delay only during perestroika. He left no academic school after him, and maybe he could not do it. His students, from several candidates of sciences to one academician, were both lucky and unlucky. Perel’man was a brilliant example in science and taught much, but he was not a “driver” of academic or administrative growth. Academician O.I. Larichev, his first student, said at the funeral: “We say goodbye to a genius.”

Perel’man did not fit into the hard environment of the late 1980s and early 1990s. In August 1994, he turned 70 and retired in September of the same year. Handing in his last paper to Avtomatika i Telemekhanika, Perel’man said that his works were finished and he was unlikely to write anymore. In late October 1994, he had to leave for the USA. Perel’man did not want to. He understood that he would not be able to work there. Unfortunately, he died 10 days before the departure.

Perel’man had worked at IARC for 38 years. Without self-delusion and ill-minded nostalgia, he loved the Institute in its best years. Perel’man will remain in IARC’s history.

Perel’man’s main books are brochures are as follows:

  1. Metodologiya primeneniya statisticheskogo modelirovaniya dlya analiza i sinteza algoritmov upravleniya (A Methodology to Apply Statistical Modeling for the Analysis and Design of Control Algorithms), Moscow: Institute of Control Sciences, the USSR Academy of Sciences, 1990. — 31 p. (coauthors N.A. Usievich and M.I. Albogachiev);
  2. Operativnaya identifikatsiya ob"ektov upravleniya (Rapid Identification of Controlled Objects), Moscow: Energoizdat, 1982. — 272 p.;
  3. Upravlenie kvaziregulyarnymi ob"ektami. Adaptivnye algoritmy tekushchei identifikatsii (Control of Quasi-Regular Objects. Adaptive Algorithms for Current Identification), Moscow: Institute of Control Sciences, the USSR Academy of Sciences, 1978. — 78 p.;
  4. Konstruirovanie i predvaritel’naya proverka ekonomicheski effektivnykh SAU (Design and Preliminary Testing of Cost-Efficient Automatic Control Systems), Moscow: Institute of Control Sciences, the USSR Academy of Sciences, 1976. — 65 p.;
  5. Upravlenie kvaziregulyarnymi ob"ektami. Modeli regressionnykh ob"ektov (Control of Quasi-Regular Objects. Models of Regression Objects), Moscow: Institute of Control Sciences, the USSR Academy of Sciences, 1975. — 77 p.;
  6. Modeli regressionnykh ob"ektov (Models of Regression Objects), Moscow: Institute of Control Sciences, the USSR Academy of Sciences, 1975. — 78 p.;
  7. Upravlenie kvaziregulyarnymi ob"ektami. Tekushchie modeli ob"ektov i metody ikh ispol’zovaniya v zadachakh upravleniya (Control of Quasi-Regular Objects. Current Models of Objects and Methods to Apply Them in Control Problems), Moscow: Institute of Control Sciences, the USSR Academy of Sciences, 1973. — 70 p.;
  8. Metody suboptimal’nogo zamknutogo upravleniya v ASU TP (Suboptimal Closed Loop Control Methods in Process Control Systems), Moscow: Institute of Control Sciences, the USSR Academy of Sciences, 1962. — 40 p.

They are presented in the Institute’s database:Перельман

The list of journal papers by Perel’man is available at Math-Net.Ru:

1. I. I. Perel’man, Object Output-Response Predictors (OORP). II. OORP Design with Incomplete a Priori Information on an Object, Avtomat. i Telemekh., 1995, 12,  3—15      Autom. Remote Control56:12 (1995), 1663—1673.
2. I. I. Perel’man, Plant Output Response Predictors (PORP). I. Methods for the Application of PORP, Avtomat. i Telemekh., 1994, 12,  3—22      Autom. Remote Control55:12 (1994), 1707—1722.
3. I. I. Perel’man, Analysis of Search Adaptation Algorithms in Closed Linear Automatic Control Systems, Avtomat. i Telemekh., 1993, 7,  110—129      Autom. Remote Control54:7 (1993), 1128—1146.
4. I. I. Perel’man, On the Practical Realizability of Methods for the Identification of a Linear Dynamic Object from Data of Its Operation in a Closed Control System, Avtomat. i Telemekh., 1992, 1,  72—86      Autom. Remote Control53:1 (1992), 59—70.
5. I. I. Perel’man, Analysis of Current Adaptive Control Methods from the Standpoint of the Application to the Automation of Technological Processes, Avtomat. i Telemekh., 1991, 7,  3—32      Autom. Remote Control52:7 (1991), 891—911.
6. I. I. Perel’man, A Stationary Adaptive Search Strategy as an Alternative Way of Stochastic Approximation, Avtomat. i Telemekh., 1987, 11,  132—143.  
7. I. I. Perel’man, Experimental Design in Development of Process Models, Avtomat. i Telemekh., 1987, 9,  3—25.  
8. I. I. Perel’man, Suboptimal Closed-Loop Stochastic Control for Minimization of the Current Production Cost, Avtomat. i Telemekh., 1984, 5,  134—143      Autom. Remote Control45:5 (1984), 663—671.
9. I. I. Perel’man, A Methodology of Choosing the Model Structure in Process Identification, Avtomat. i Telemekh., 1983, 11,  5—29      Autom. Remote Control44:11 (1983), 1389—1408.
10. I. I. Perel’man, Methods for Sound Estimation of Linear Dynamic Plant Parameters and Feasibility of Their Implementation on Finite Samples, Avtomat. i Telemekh., 1981, 3,  49—55      Autom. Remote Control42:3 (1981), 309—313.
11. I. I. Perel’man, Dynamic Optimization in Control Systems by Conditional Forecasting Algorithms, Avtomat. i Telemekh., 1978, 9,  146—160.  
12. I. I. Perel’man, An Adaptive Approach tо Data Weighting in Estimation of Unobservable Drifting Parameters, Avtomat. i Telemekh., 1977, 4,  88—100    Autom. Remote Control38:4 (1977), 524—535.
13. N. M. Emel’yanova, I. I. Perel’man, The Problem of Conditional Interval Forecasting and Its Solution by Using a Generalized Plant Model, Avtomat. i Telemekh., 1975, 11,  64—79.  
14. I. I. Perel’man, Asymptotic Properties of Regression Models of Controlled Plants, Avtomat. i Telemekh., 1975, 4,  56—63    Autom. Remote Control36:4 (1975), 572—579.

The list of his papers in Avtomatika i Telemekhanika can be found at:

For their English versions, see the microfilm collection of Automation and Remote Control:

Many inventions by Perel’man are available at:

Scopus Author ID: 7003360316