General Constraints Research Articles

Randomized Parallel Algorithm for Maximizing Nonsubmodular Function Subject to Cardinality Constraint

The problem of maximization submodular functions with a cardinality constraint has been extensively researched in recent years. Balkanski and Singer were the first to study this class problem. Subsequently, Chekuri and Kent recently extended these results to more general constraints, that is, [Formula: see text]-cardinality constraint, partition and laminar matroids, matching, knapsack constraints, and including their intersection. They proposed a [Formula: see text] approximation randomized-parallel-greedy algorithm which are poly-logarithmic adaptivity. However, these existing approaches are hardly extended to the nonsubmodular case. In this paper, we investigate the problem of maximization on a nonsubmodular function subject to a cardinality constraint, provided the objective function is specified by a generic submodularity ratio [Formula: see text]. We design a [Formula: see text]-approximation Greedy algorithm by using the technical aspects to maximize the multilinear relaxation of the object function under the [Formula: see text]-cardinality constraints. The adaptive of multilinear relaxation is [Formula: see text]; the number of oracle is [Formula: see text].

Direct dynamic-simulation approach to trajectory optimization

This paper proposes a new direct method for an efficient trajectory optimization using the point that the dynamics of a deterministic system are uniquely determined by initial states and controls imposed over the time horizon of interest. To effectively implement this concept, the Hermite spline is adopted to interpolate the continuous controls and the system dynamics are integrated with corresponding control parameters in prior. As a result, the optimal control problem can be transcribed into a nonlinear programming problem which has no dynamic equality constraints and no intermediate states in its design variables. In addition, the paper proposes an efficient recursive Jacobian estimation technique and introduces a Jacobian transformation matrix to straightforwardly handle the general state constraints. Important properties of the present method are thoroughly investigated through its applications to the trajectory optimization for a soft lunar landing from a parking orbit, including the detailed analyses for the de-orbiting phase. The computed results are compared with those using the pseudo-spectral method to demonstrate an extreme outperformance of the proposed method in the aerospace applications over the traditional direct method.

Beyond Hammersley’s Last-Passage Percolation: a discussion on possible local and global constraints

Hammersley’s Last-Passage Percolation (LPP), also known as Ulam’s problem, is a well-studied model that can be described as follows: let m points be chosen uniformly and independently in [0,1] ^2 , then what is the maximal number \mathcal L_m of points that can be collected by an up-right path? We introduce here a generalization of this LPP, allowing for more general constraints than the up-right condition: the constraints may be either local or global . We give the correct order of \mathcal L_m in a general manner, and we illustrate the interest and usefulness of this generalized LPP with examples and simulations.

Semi-martingale driven variational principles

Spearheaded by the recent efforts to derive stochastic geophysical fluid dynamics models, we present a general framework for introducing stochasticity into variational principles through the concept of a semi-martingale driven variational principle and constraining the component variables to be compatible with the driving semi-martingale. Within this framework and the corresponding choice of constraints, the Euler–Poincaré equation can be easily deduced. We show that the deterministic theory is a special case of this class of stochastic variational principles. Moreover, this is a natural framework that enables us to correctly characterize the pressure term in incompressible stochastic fluid models. Other general constraints can also be incorporated as long as they are compatible with the driving semi-martingale.

Maximum Entropy on the Mean approach to solve generalized inverse problems with an application in computational thermodynamics

In this paper, we study entropy maximisation problems in order to reconstruct functions or measures subject to very general integral constraints. Our work has a twofold purpose. We first make a global synthesis of entropy maximisation problems in the case of a single reconstruction (measure or function) from the convex analysis point of view, as well as in the framework of the embedding into the Maximum Entropy on the Mean (MEM) setting. We further propose an extension of the entropy methods for a multidimensional case.

A fast saddle-point dynamical system approach to robust deep learning

Recent focus on robustness to adversarial attacks for deep neural networks produced a large variety of algorithms for training robust models. Most of the effective algorithms involve solving the min–max optimization problem for training robust models (min step) under worst-case attacks (max step). However, they often suffer from high computational cost from running several inner maximization iterations (to find an optimal attack) inside every outer minimization iteration. Therefore, it becomes difficult to readily apply such algorithms for moderate to large size real world data sets. To alleviate this, we explore the effectiveness of iterative descent–ascent algorithms where the maximization and minimization steps are executed in an alternate fashion to simultaneously obtain the worst-case attack and the corresponding robust model. Specifically, we propose a novel discrete-time dynamical system-based algorithm that aims to find the saddle point of a min–max optimization problem in the presence of uncertainties. Under the assumptions that the cost function is convex and uncertainties enter concavely in the robust learning problem, we analytically show that our algorithm converges asymptotically to the robust optimal solution under a general adversarial budget constraints as induced by ℓp norm, for 1≤p≤∞. Based on our proposed analysis, we devise a fast robust training algorithm for deep neural networks. Although such training involves highly non-convex robust optimization problems, empirical results show that the algorithm can achieve significant robustness compared to other state-of-the-art robust models on benchmark data sets.

Improved variance estimation for inequality-constrained domain mean estimators using survey data

In survey domain estimation, a priori information can often be imposed in the form of linear inequality constraints on the domain estimators. Wu et al. (2016) formulated the isotonic domain mean estimator, for the simple order restriction, and methods for more general constraints were proposed in Oliva-Avilés et al. (2020). When the assumptions are valid, imposing restrictions on the estimators will ensure that the a priori information is respected, and in addition allows information to be pooled across domains, resulting in estimators with smaller variance. Here, we propose a method to further improve the estimation of the covariance matrix for these constrained domain estimators, using a mixture of possible covariance matrices obtained from the inequality constraints. We prove consistency of the improved variance estimator, and simulations demonstrate that the new estimator results in improved coverage probabilities for domain mean confidence intervals, while retaining the smaller confidence interval lengths.

Relations between Abs-Normal NLPs and MPCCs. Part 1: Strong Constraint Qualifications

This work is part of an ongoing effort of comparing non-smooth optimization problems in abs-normal form to MPCCs. We study the general abs-normal NLP with equality and inequality constraints in relation to an equivalent MPCC reformulation. We show that kink qualifications and MPCC constraint qualifications of linear independence type and Mangasarian-Fromovitz type are equivalent. Then we consider strong stationarity concepts with first and second order optimality conditions, which again turn out to be equivalent for the two problem classes. Throughout we also consider specific slack reformulations suggested in [9], which preserve constraint qualifications of linear independence type but not of Mangasarian-Fromovitz type.

A Novel Multiagent Neurodynamic Approach to Constrained Distributed Convex Optimization.

This paper considers a class of distributed convex optimization problems with constraints and gives a novel multiagent neurodynamic approach in continuous-time form. The considered distributed optimization is to search for a minimizer of the summation of nonsmooth convex functions on some agents, which have local general constraints. The proposed approach solves the objective function of each agent individually, and the state solutions of all agents reach consensus asymptotically under mild assumptions. In particular, the existence and boundedness of the global state solution to the dynamical system are guaranteed. Moreover, the state solution reaches the feasible region of equivalent optimization problem asymptotically and the output of each agent is convergent to the optimal solution set of the primal distributed problem. In contrast to the existing methods in a distributed manner, the proposed approach is more convenient for general constrained distributed problems and has low structure complexity which could narrow the bandwidth of communication. Finally, the proposed neurodynamic approach is applied to two numerical examples and a class of power system optimal load-sharing problems to support the theoretical results and its efficiency.

A penalty method for nonlinear programs with set exclusion constraints

A common requirement in optimal control problems arising in autonomous navigation is that the decision variables are constrained to be outside certain sets. Such set exclusion constraints represent obstacles that must be avoided by the motion system. This paper presents a simple and efficient method for solving optimization problems with general set exclusion and implicit constraints. The method embeds the set exclusion constraints in a quadratic penalty framework and solves the inner optimization problems using a proximal algorithm that deals directly with the implicit constraints. We derive convergence results for this method by transforming the generated iterates to points of a reformulated problem with complementarity constraints. Furthermore, the practical application of the solution method is validated in numerical simulations of a model predictive control approach to path planning for a mobile robot. Finally, a runtime comparison with state-of-the-art solvers applied to the problem with complementarity constraints illustrates the efficiency of the proposed method.

Probing fundamental physics with gravitational waves: The next generation

Gravitational wave observations of compact binary mergers are already providing stringent tests of general relativity and constraints on modified gravity. Ground-based interferometric detectors will soon reach design sensitivity and they will be followed by third-generation upgrades, possibly operating in conjunction with space-based detectors. How will these improvements affect our ability to investigate fundamental physics with gravitational waves? The answer depends on the timeline for the sensitivity upgrades of the instruments, but also on astrophysical compact binary population uncertainties, which determine the number and signal-to-noise ratio of the observed sources. We consider several scenarios for the proposed timeline of detector upgrades and various astrophysical population models. Using a stacked Fisher matrix analysis of binary black hole merger observations, we thoroughly investigate future theory-agnostic bounds on modifications of general relativity, as well as bounds on specific theories. For theory-agnostic bounds, we find that ground-based observations of stellar-mass black holes and LISA observations of massive black holes can each lead to improvements of 2-4 orders of magnitude with respect to present gravitational wave constraints, while multiband observations can yield improvements of 1-6 orders of magnitude. We also clarify how the relation between theory-agnostic and theory-specific bounds depends on the source properties.

Welfare-maximizing Guaranteed Dashboard Mechanisms

Bidding dashboards are used in online marketplaces to aid a bidder in computing good bidding strategies, particularly when the auction used by the marketplace is constrained to have the winners-pay-bid payment format. A dashboard predicts the outcome a bidder can expect to get at each possible bid. To convince a bidder to best respond to the information published in a dashboard, a dashboard mechanism should ensure either (a) that best responding maximizes the bidder's utility (a weaker requirement) or (b) that the mechanism implements the outcome published in the dashboard (a stronger requirement that subsumes (a)). Recent work by Hartline et al. EC'19 formalized the notion of dashboard mechanisms and designed winners-pay-bid mechanisms that guaranteed epsilon-optimal utility (an epsilon-approximate version of (a)), but not (b). I.e., the mechanism could end up implementing arbitrarily different outcomes from what was promised. While this guarantee is sufficient from a purely technical perspective, it is far from enough in the real world: it is hard to convince bidders to best respond to information which could be arbitrarily inaccurate, regardless of the theoretical promise of near-optimality. In this paper we study guaranteed dashboard mechanisms, namely, ones that are guaranteed to implement what they publish, and obtain good welfare. We study this question in a repeated auction setting for general single-dimensional valuations and give tight characterizations of the loss in welfare as a function of natural parameters upper bounding the difference in valuation profile across the rounds. In particular, we give three different characterizations, bounding the loss in welfare in terms of the 0 norm, 1 norm and infinite norm of difference in valuation profile across rounds. All the characterizations generalize at least up to matroid feasibility constraints, and the infinite norm characterization extends to general downward-closed feasibility constraints. We bring to bear different techniques for each of these characterizations, including connections to differential privacy and online convex optimizations.

Plant’s gypsum affinity shapes responses to specific edaphic constraints without limiting responses to other general constraints

Harsh edaphic environments harbor species with different soil affinities. Plant’s responses to specific edaphic constraints may be compromised against responses to prevalent stresses shared with other semi-arid environments. We expect that species with high edaphic affinity may show traits to overcome harsh soil properties, while species with low affinity may respond to environmental constraints shared with arid environments. We quantified the edaphic affinity of 12 plant species co-occurring in gypsum outcrops and measured traits related to plant responses to specific gypsum constraints (rooting and water uptake depth, foliar accumulation of Ca, S and Mg), and traits related to common constraints of arid environments (water use efficiency, macronutrients foliar content). Plants in gypsum outcrops differed in their strategies to face edaphic limitations. A phylogenetic informed PCA segregated species based on their foliar Ca and S accumulation and greater water uptake depths, associated with plant responses to specific gypsum limitations. Species’ gypsum affinity explained this segregation, but traits related to water or nutrient use efficiency did not contribute substantially to this axis. Plant’s specializations to respond to specific edaphic constraints of gypsum soils do not limit their ability to deal with other non-specific environmental constraints.

A Service System with Packing Constraints: Greedy Randomized Algorithm Achieving Sublinear in Scale Optimality Gap

A service system with multiple types of arriving customers is considered. There is an infinite number of homogeneous servers. Multiple customers can be placed for simultaneous service into one server, subject to general packing constraints. The service times of different customers are independent even if they are served simultaneously by the same server; the service time distribution depends on the customer type. Each new arriving customer is placed for service immediately into either an occupied server, that is, one already serving other customers, as long as packing constraints are not violated or into an empty server. After service completion, each customer leaves its server and the system. The basic objective is to minimize the number of occupied servers in steady state. We study a greedy random (GRAND) placement (packing) algorithm, introduced in our previous work. This is a simple online algorithm that places each arriving customer uniformly at random into either one of the already occupied servers that can still fit the customer or one of the so-called zero servers, which are empty servers designated to be available to new arrivals. In our previous work, a version of the algorithm, labeled GRAND(aZ), is considered, in which the number of zero servers is aZ with Z being the current total number of customers in the system and positive a being an algorithm parameter. GRAND(aZ) is shown in our previous work to be asymptotically optimal in the following sense: (a) the steady-state optimality gap grows linearly in the system scale r (the mean total number of customers in service), that is, as c(a)r for some positive c(a), and (b) c(a) vanishes as a goes to zero. In this paper, we consider the GRAND(Z p ) algorithm, in which the number of zero servers is Z p , where p < 1 is a fixed parameter, sufficiently close to 1. We prove the asymptotic optimality of GRAND(Z p ) in the sense that the steady-state optimality gap is sublinear in the system scale r. This is a stronger form of asymptotic optimality than that of GRAND(aZ).

Mathcal{N} = 2 AdS4 supergravity, holography and Ward identities

We develop in detail the holographic framework for an mathcal{N} = 2 pure AdS supergravity model in four dimensions, including all the contributions from the fermionic fields and adopting the Fefferman-Graham parametrization. We work in the first order formalism, where the full superconformal structure can be kept manifest in principle, even if only a part of it is realized as a symmetry on the boundary, while the remainder has a non-linear realization. Our study generalizes the results presented in antecedent literature and includes a general discussion of the gauge-fixing conditions on the bulk fields which yield the asymptotic symmetries at the boundary. We construct the corresponding super- conformal currents and show that they satisfy the related Ward identities when the bulk equations of motion are imposed. Consistency of the holographic setup requires the super- AdS curvatures to vanish at the boundary. This determines, in particular, the expression of the super-Schouten tensor of the boundary theory, which generalizes the purely bosonic Schouten tensor of standard gravity by including gravitini bilinears. The same applies to the superpartner of the super-Schouten tensor, the conformino. Furthermore, the vanishing of the supertorsion poses general constraints on the sources of the three-dimensional boundary conformal field theory and requires that the super-Schouten tensor is endowed with an antisymmetric part proportional to a gravitino-squared term.

C-lasso - a Python package for constrained sparse and robust regression and classification

We introduce c-lasso, a Python package that enables sparse and robust linear regression and classification with linear equality constraints. The underlying statistical forward model is assumed to be of the following form: \[ y = X \beta + \sigma \epsilon \qquad \textrm{subject to} \qquad C\beta=0 \] Here, $X \in \mathbb{R}^{n\times d}$is a given design matrix and the vector $y \in \mathbb{R}^{n}$ is a continuous or binary response vector. The matrix $C$ is a general constraint matrix. The vector $\beta \in \mathbb{R}^{d}$ contains the unknown coefficients and $\sigma$ an unknown scale. Prominent use cases are (sparse) log-contrast regression with compositional data $X$, requiring the constraint $1_d^T \beta = 0$ (Aitchion and Bacon-Shone 1984) and the Generalized Lasso which is a special case of the described problem (see, e.g, (James, Paulson, and Rusmevichientong 2020), Example 3). The c-lasso package provides estimators for inferring unknown coefficients and scale (i.e., perspective M-estimators (Combettes and Muller 2020a)) of the form \[ \min_{\beta \in \mathbb{R}^d, \sigma \in \mathbb{R}_{0}} f\left(X\beta - y,{\sigma} \right) + \lambda \left\lVert \beta\right\rVert_1 \qquad \textrm{subject to} \qquad C\beta = 0 \] for several convex loss functions $f(\cdot,\cdot)$. This includes the constrained Lasso, the constrained scaled Lasso, and sparse Huber M-estimators with linear equality constraints.

Structural Synthesis of Lower-Class Robot Manipulators with General Constraint One

In this research, the structural synthesis of lower-class robot manipulators with general constraint one is investigated. The number of constraints d = 1 and the mobility number λ = 5 for sphere–sphere and sphere–plane combinations of higher kinematic pairs are investigated by using the analytical approach. Structural synthesis of new first- and second-class robot manipulators is investigated by using transformations of higher kinematic pairs to lower kinematic pairs. Therefore, using this approach, we obtain structural groups with general constraint one. Further, kinematic structures and physical models of overconstrained robot manipulators are presented in the research.

Cheaper relaxation and better approximation for multi-ball constrained quadratic optimization and extension

We propose a convex quadratic programming (CQP) relaxation for multi-ball constrained quadratic optimization (MB). (CQP) is shown to be equivalent to semidefinite programming relaxation in the hard case. Based on (CQP), we propose an algorithm for solving (MB), which returns a solution of (MB) with an approximation bound independent of the number of constraints. The approximation algorithm is further extended to solve nonconvex quadratic optimization with more general constraints. As an application, we propose a standard quadratic programming relaxation for finding Chebyshev center of a general convex set with a guaranteed approximation bound.

Lexicographic multi-objective MPC for constrained nonlinear systems with changing objective prioritization

This paper proposes a new lexicographic multi-objective model predictive control (MoMPC) scheme for constrained nonlinear systems with changing objective prioritization. The lexicographic optimization-based receding horizon optimal control problem is firstly formulated to accommodate the concern on prioritized control of multiple (economic) objectives. Then an elaborate generic function and the notion of general terminal constraints are introduced to regain feasibility of lexicographic MoMPC in the presence of constraints and changing objective prioritization. Moreover, the LaSalle invariance set theorem is used to establish stability of the closed-loop system with the lexicographic MoMPC. An example of a coupled-tank level control problem with three conflicting objectives is adopted to demonstrate the effectiveness of the proposed scheme.

Unified IRS-Aided MIMO Transceiver Designs via Majorization Theory

In this paper, we develop a unified framework for IRS-aided transceiver designs under general power constraints in multiple-input multiple-output (MIMO) systems which implement interference (pre-)subtraction via Tomlinson-Harashima precoding (THP) or Decision Feedback Equalization (DFE) technologies. Armed with majorization theory, two fundamental classes of performance criteria, namely $K$ -increasing Schur-concave and Schur-convex functions of the logarithm of Mean Square Error (MSE) of the data stream, are investigated in depth. Firstly, we propose a simplified counterpart of the optimal transceiver design under general power constraints, with equivalence guaranteed by Pareto optimization theory and Lagrange duality. Moreover, the optimal semi-closed form solution to this simplified transceiver design can be attained using the modified subgradient method. Next, we prove that for any Schur-concave objective, the optimal nonlinear THP (DFE) design is in essence the linear precoding (equalization). For any Schur-convex objective, the optimal transceiver design results in individual data streams with equal MSEs, and thereby reduces to the Gaussian mutual information maximization based design. Based on the above conclusions, we further propose an efficient alternating optimization algorithm to decouple the optimization of the transmit precoder and the IRS reflection coefficients, where the classical successive convex approximation (SCA) technique is applied to fight against non-convex subproblems. From the low computational complexity perspective, a two-stage scheme is also developed inspired by the capability of the IRS in constructing favorable wireless links. Finally, numerical results show the global optimality of the modified subgradient method and the excellent performance of the proposed alternating optimization algorithm and two-stage scheme.