All issues

In this paper, we consider convex stochastic optimization problems arising in machine learning applications (e. g., risk minimization) and mathematical statistics (e. g., maximum likelihood estimation). There are two main approaches to solve such kinds of problems, namely the Stochastic Approximation approach (online approach) and the Sample Average Approximation approach, also known as the Monte Carlo approach, (offline approach). In the offline approach, the problem is replaced by its empirical counterpart (the empirical risk minimization problem). The natural question is how to define the problem sample size, i. e., how many realizations should be sampled so that the quite accurate solution of the empirical problem be the solution of the original problem with the desired precision. This issue is one of the main issues in modern machine learning and optimization. In the last decade, a lot of significant advances were made in these areas to solve convex stochastic optimization problems on the Euclidean balls (or the whole space). In this work, we are based on these advances and study the case of arbitrary balls in the $p$-norms. We also explore the question of how the parameter $p$ affects the estimates of the required number of terms as a function of empirical risk.

In this paper, both convex and saddle point optimization problems are considered. For strongly convex problems, the existing results on the same sample sizes in both approaches (online and offline) were generalized to arbitrary norms. Moreover, it was shown that the strong convexity condition can be weakened: the obtained results are valid for functions satisfying the quadratic growth condition. In the case when this condition is not met, it is proposed to use the regularization of the original problem in an arbitrary norm. In contradistinction to convex problems, saddle point problems are much less studied. For saddle point problems, the sample size was obtained under the condition of $\gamma$-growth of the objective function. When $\gamma = 1$, this condition is the condition of sharp minimum in convex problems. In this article, it was shown that the sample size in the case of a sharp minimum is almost independent of the desired accuracy of the solution of the original problem.

In this paper, we develop a new first-order method for composite nonconvex minimization problems with simple constraints and inexact oracle. The objective function is given as a sum of «hard», possibly nonconvex part, and «simple» convex part. Informally speaking, oracle inexactness means that, for the «hard» part, at any point we can approximately calculate the value of the function and construct a quadratic function, which approximately bounds this function from above. We give several examples of such inexactness: smooth nonconvex functions with inexact H¨older-continuous gradient, functions given by the auxiliary uniformly concave maximization problem, which can be solved only approximately. For the introduced class of problems, we propose a gradient-type method, which allows one to use a different proximal setup to adapt to the geometry of the feasible set, adaptively chooses controlled oracle error, allows for inexact proximal mapping. We provide a convergence rate for our method in terms of the norm of generalized gradient mapping and show that, in the case of an inexact Hölder-continuous gradient, our method is universal with respect to Hölder parameters of the problem. Finally, in a particular case, we show that the small value of the norm of generalized gradient mapping at a point means that a necessary condition of local minimum approximately holds at that point.

This article reviews both historical achievements and modern results in the field of Markov Decision Process (MDP) and convex optimization. This review is the first attempt to cover the field of reinforcement learning in Russian in the context of convex optimization. The fundamental Bellman equation and the criteria of optimality of policy — strategies based on it, which make decisions based on the known state of the environment at the moment, are considered. The main iterative algorithms of policy optimization based on the solution of the Bellman equations are also considered. An important section of this article was the consideration of an alternative to the $Q$-learning approach — the method of direct maximization of the agent’s average reward for the chosen strategy from interaction with the environment. Thus, the solution of this convex optimization problem can be represented as a linear programming problem. The paper demonstrates how the convex optimization apparatus is used to solve the problem of Reinforcement Learning (RL). In particular, it is shown how the concept of strong duality allows us to naturally modify the formulation of the RL problem, showing the equivalence between maximizing the agent’s reward and finding his optimal strategy. The paper also discusses the complexity of MDP optimization with respect to the number of state–action–reward triples obtained as a result of interaction with the environment. The optimal limits of the MDP solution complexity are presented in the case of an ergodic process with an infinite horizon, as well as in the case of a non-stationary process with a finite horizon, which can be restarted several times in a row or immediately run in parallel in several threads. The review also reviews the latest results on reducing the gap between the lower and upper estimates of the complexity of MDP optimization with average remuneration (Averaged MDP, AMDP). In conclusion, the real-valued parametrization of agent policy and a class of gradient optimization methods through maximizing the $Q$-function of value are considered. In particular, a special class of MDPs with restrictions on the value of policy (Constrained Markov Decision Process, CMDP) is presented, for which a general direct-dual approach to optimization with strong duality is proposed.

As the population of cities grows, the need to plan for the development of transport infrastructure becomes more acute. For this purpose, transport modeling packages are created. These packages usually contain a set of convex optimization problems, the iterative solution of which leads to the desired equilibrium distribution of flows along the paths. One of the directions for the development of transport modeling is the construction of more accurate generalized models that take into account different types of passengers, their travel purposes, as well as the specifics of personal and public modes of transport that agents can use. Another important direction of transport models development is to improve the efficiency of the calculations performed. Since, due to the large dimension of modern transport networks, the search for a numerical solution to the problem of equilibrium distribution of flows along the paths is quite expensive. The iterative nature of the entire solution process only makes this worse. One of the approaches leading to a reduction in the number of calculations performed is the construction of consistent models that allow to combine the blocks of a 4-stage model into a single optimization problem. This makes it possible to eliminate the iterative running of blocks, moving from solving a separate optimization problem at each stage to some general problem. Early work has proven that such approaches provide equivalent solutions. However, it is worth considering the validity and interpretability of these methods. The purpose of this article is to substantiate a single problem, that combines both the calculation of the trip matrix and the modal choice, for the generalized case when there are different layers of demand, types of agents and classes of vehicles in the transport network. The article provides possible interpretations for the gauge parameters used in the problem, as well as for the dual factors associated with the balance constraints. The authors of the article also show the possibility of combining the considered problem with a block for determining network load into a single optimization problem.

Since 1950s the field of city transport modelling has progressed rapidly. The first equilibrium distribution models of traffic flow appeared. The most popular model (which is still being widely used) was the Beckmann model, based on the two Wardrop principles. The core of the model could be briefly described as the search for the Nash equilibrium in a population demand game, in which losses of agents (drivers) are calculated based on the chosen path and demands of this path with correspondences being fixed. The demands (costs) of a path are calculated as the sum of the demands of different path segments (graph edges), that are included in the path. The costs of an edge (edge travel time) are determined by the amount of traffic on this edge (more traffic means larger travel time). The flow on a graph edge is determined by the sum of flows over all paths passing through the given edge. Thus, the cost of traveling along a path is determined not only by the choice of the path, but also by the paths other drivers have chosen. Thus, it is a standard game theory task. The way cost functions are constructed allows us to narrow the search for equilibrium to solving an optimization problem (game is potential in this case). If the cost functions are monotone and non-decreasing, the optimization problem is convex. Actually, different assumptions about the cost functions form different models. The most popular model is based on the BPR cost function. Such functions are massively used in calculations of real cities. However, in the beginning of the XXI century, Yu. E. Nesterov and A. de Palma showed that Beckmann-type models have serious weak points. Those could be fixed using the stable dynamics model, as it was called by the authors. The search for equilibrium here could be also reduced to an optimization problem, moreover, the problem of linear programming. In 2013, A.V.Gasnikov discovered that the stable dynamics model can be obtained by a passage to the limit in the Beckmann model. However, it was made only for several practically important, but still special cases. Generally, the question if this passage to the limit is possible remains open. In this paper, we provide the justification of the possibility of the above-mentioned passage to the limit in the general case, when the cost function for traveling along the edge as a function of the flow along the edge degenerates into a function equal to fixed costs until the capacity is reached and it is equal to plus infinity when the capacity is exceeded.

Problems of multiple covering ($k$-covering) of a bounded set $G$ with equal circles of a given radius are well known. They are thoroughly studied under the assumption that $G$ is a finite set. There are several papers concerned with studying this problem in the case where $G$ is a connected set. In this paper, we study the problem of minimizing the number of circles that form a $k$-covering, $k \geqslant 1$, provided that $G$ is a bounded convex plane domain.

For the above-mentioned problem, we state a 0-1 linear model, a general integer linear model, and a nonlinear model, imposing a constraint on the minimum distance between the centers of covering circles. The latter constraint is due to the fact that in practice one can place at most one device at each point. We establish necessary and sufficient solvability conditions for the linear models and describe one (easily realizable) variant of these conditions in the case where the covered set $G$ is a rectangle.

We propose some methods for finding an approximate number of circles of a given radius that provide the desired $k$-covering of the set $G$, both with and without constraints on distances between the circles’ centers. We treat the calculated values as approximate upper bounds for the number of circles. We also propose a technique that allows one to get approximate lower bounds for the number of circles that is necessary for providing a $k$-covering of the set $G$. In the general linear model, as distinct from the 0-1 linear model, we require no additional constraint. The difference between the upper and lower bounds for the number of circles characterizes the quality (acceptability) of the constructed $k$-covering.

We state a nonlinear mathematical model for the $k$-covering problem with the above-mentioned constraints imposed on distances between the centers of covering circles. For this model, we propose an algorithm which (in certain cases) allows one to find more exact solutions to covering problems than those calculated from linear models.

For implementing the proposed approach, we have developed computer programs and performed numerical experiments. Results of numerical experiments demonstrate the effectiveness of the method.

In this paper we propose high-order (tensor) methods for two types of saddle point problems. Firstly, we consider the classic min-max saddle point problem. Secondly, we consider the search for a stationary point of the saddle point problem objective by its gradient norm minimization. Obviously, the stationary point does not always coincide with the optimal point. However, if we have a linear optimization problem with linear constraints, the algorithm for gradient norm minimization becomes useful. In this case we can reconstruct the solution of the optimization problem of a primal function from the solution of gradient norm minimization of dual function. In this paper we consider both types of problems with no constraints. Additionally, we assume that the objective function is $\mu$-strongly convex by the first argument, $\mu$-strongly concave by the second argument, and that the $p$-th derivative of the objective is Lipschitz-continous.

For min-max problems we propose two algorithms. Since we consider strongly convex a strongly concave problem, the first algorithm uses the existing tensor method for regular convex concave saddle point problems and accelerates it with the restarts technique. The complexity of such an algorithm is linear. If we additionally assume that our objective is first and second order Lipschitz, we can improve its performance even more. To do this, we can switch to another existing algorithm in its area of quadratic convergence. Thus, we get the second algorithm, which has a global linear convergence rate and a local quadratic convergence rate.

Finally, in convex optimization there exists a special methodology to solve gradient norm minimization problems by tensor methods. Its main idea is to use existing (near-)optimal algorithms inside a special framework. I want to emphasize that inside this framework we do not necessarily need the assumptions of strong convexity, because we can regularize the convex objective in a special way to make it strongly convex. In our article we transfer this framework on convex-concave objective functions and use it with our aforementioned algorithm with a global linear convergence and a local quadratic convergence rate.

Since the saddle point problem is a particular case of the monotone variation inequality problem, the proposed methods will also work in solving strongly monotone variational inequality problems.

Stochastic optimization is a current area of research due to significant advances in machine learning and their applications to everyday problems. In this paper, we consider two fundamentally different methods for solving the problem of stochastic optimization — online and offline algorithms. The corresponding algorithms have their qualitative advantages over each other. So, for offline algorithms, it is required to solve an auxiliary problem with high accuracy. However, this can be done in a distributed manner, and this opens up fundamental possibilities such as, for example, the construction of a dual problem. Despite this, both online and offline algorithms pursue a common goal — solving the stochastic optimization problem with a given accuracy. This is reflected in the comparison of the computational complexity of the described algorithms, which is demonstrated in this paper.

The comparison of the described methods is carried out for two types of stochastic problems — convex optimization and saddles. For problems of stochastic convex optimization, the existing solutions make it possible to compare online and offline algorithms in some detail. In particular, for strongly convex problems, the computational complexity of the algorithms is the same, and the condition of strong convexity can be weakened to the condition of $\gamma$-growth of the objective function. From this point of view, saddle point problems are much less studied. Nevertheless, existing solutions allow us to outline the main directions of research. Thus, significant progress has been made for bilinear saddle point problems using online algorithms. Offline algorithms are represented by just one study. In this paper, this example demonstrates the similarity of both algorithms with convex optimization. The issue of the accuracy of solving the auxiliary problem for saddles was also worked out. On the other hand, the saddle point problem of stochastic optimization generalizes the convex one, that is, it is its logical continuation. This is manifested in the fact that existing results from convex optimization can be transferred to saddles. In this paper, such a transfer is carried out for the results of the online algorithm in the convex case, when the objective function satisfies the $\gamma$-growth condition.

The Lipschitz continuous property has been used for a long time to solve the global optimization problem and continues to be used. Here we can mention the work of Piyavskii, Yevtushenko, Strongin, Shubert, Sergeyev, Kvasov and others. Most papers assume a priori knowledge of the Lipschitz constant, but the derivation of this constant is a separate problem. Further still, we must prove that an objective function is really Lipschitz, and it is a complicated problem too. In the case where the Lipschitz continuity is established, Strongin proposed an algorithm for global optimization of a satisfying Lipschitz condition on a compact interval function without any a priori knowledge of the Lipschitz estimate. The algorithm not only finds a global extremum, but it determines the Lipschitz estimate too. It is known that every function that satisfies the Lipchitz condition on a compact convex set is uniformly continuous, but the reverse is not always true. However, there exist models (Arutyunova, Dulliev, Zabotin) whose study requires a minimization of the continuous but definitely not Lipschitz function. One of the algorithms for solving such a problem was proposed by R. J. Vanderbei. In his work he introduced some generalization of the Lipchitz property named $\varepsilon$-Lipchitz and proved that a function defined on a compact convex set is uniformly continuous if and only if it satisfies the $\varepsilon$-Lipchitz condition. The above-mentioned property allowed him to extend Piyavskii’s method. However, Vanderbei assumed that for a given value of $\varepsilon$ it is possible to obtain an associate Lipschitz $\varepsilon$-constant, which is a very difficult problem. Thus, there is a need to construct, for a function continuous on a compact convex domain, a global optimization algorithm which works in some way like Strongin’s algorithm, i.e., without any a priori knowledge of the Lipschitz $\varepsilon$-constant. In this paper we propose an extension of Strongin’s global optimization algorithm to a function continuous on a compact interval using the $\varepsilon$-Lipchitz conception, prove its convergence and solve some numerical examples using the software that implements the developed method.

In this article we consider min-min type of problems or minimization by two groups of variables. In some way it is similar to classic min-max saddle point problem. Although, saddle point problems are usually more difficult in some way. Min-min problems may occur in case if some groups of variables in convex optimization have different dimensions or if these groups have different domains. Such problem structure gives us an ability to split the main task to subproblems, and allows to tackle it with mixed oracles. However existing articles on this topic cover only zeroth and first order oracles, in our work we consider high-order tensor methods to solve inner problem and fast gradient method to solve outer problem.

We assume, that outer problem is constrained to some convex compact set, and for the inner problem we consider both unconstrained case and being constrained to some convex compact set. By definition, tensor methods use high-order derivatives, so the time per single iteration of the method depends a lot on the dimensionality of the problem it solves. Therefore, we suggest, that the dimension of the inner problem variable is not greater than 1000. Additionally, we need some specific assumptions to be able to use mixed oracles. Firstly, we assume, that the objective is convex in both groups of variables and its gradient by both variables is Lipschitz continuous. Secondly, we assume the inner problem is strongly convex and its gradient is Lipschitz continuous. Also, since we are going to use tensor methods for inner problem, we need it to be p-th order Lipschitz continuous ($p > 1$). Finally, we assume strong convexity of the outer problem to be able to use fast gradient method for strongly convex functions.

We need to emphasize, that we use superfast tensor method to tackle inner subproblem in unconstrained case. And when we solve inner problem on compact set, we use accelerated high-order composite proximal method.

Additionally, in the end of the article we compare the theoretical complexity of obtained methods with regular gradient method, which solves the mentioned problem as regular convex optimization problem and doesn’t take into account its structure (Remarks 1 and 2).