Applied and Computational Mathematics
Volume 5, Issue 5, October 2016, Pages: 213-229

Distributed Subgradient Algorithm for Multi-Agent Convex Optimization with Global Inequality and Equality Constraints

Li Xiao*, Junjie Bao, Xi Shi

Department of Mathematics and Information Engineering, Chongqing University of Education, Chongqing, PR China

Email address:

(Li Xiao)

*Corresponding author

To cite this article:

Li Xiao, Junjie Bao, Xi Shi. Distributed Subgradient Algorithm for Multi-Agent Convex Optimization with Global Inequality and Equality Constraints. Applied and Computational Mathematics. Vol. 5, No. 5, 2016, pp. 213-229. doi: 10.11648/j.acm.20160505.15

Received: August 7, 2016; Accepted: October 5, 2016; Published: October 27, 2016

Abstract: In this paper, we present an improved subgradient algorithm for solving a general multi-agent convex optimization problem in a distributed way,where the agents are to jointly minimize a global objective function subject to a global inequality constraint, a global equality constraint and a global constraint set. The global objective function is a combination of local agent objective functions and the global constraint set isthe intersection of each agent local constraint set. Our motivation comes from networking applications where dual and primal-dual subgradient methods have attracted much attention in the design of decentralized network protocols. Our main focus is on constrained problems where the local constraint sets are identical. Thus, we propose a distributed primal-dual subgradient algorithm, which isbased on the description of the primal-dual optimal solutions as the saddle points of the penalty functions. We show that,the algorithm can be implemented over networks with changing topologies but satisfying a standard connectivity property, and allow the agents to asymptotically converge to optimal solution with optimal value of the optimization problem under the Slater’s condition.

Keywords: Consensus, Saddle Point, Distributed Optimization, Subgradient Algorithm

1. Introduction

In recent years, distributed optimization and control have developed rapidly, and have been welcomed in the fields of industry and national defense, including smart grid, sensor network, social network and information system (Cyber-Physical system). Distributed optimization problems of multi-agent systems appear different kinds of distributed processing issues such as distributed estimation, distributed motion planning, distributed resource allocation and distributed congestion control [1-12]. The main focus is to solve a distributed optimization problem where the global objective function is composed of a sum of local objective functions, each of which is only known by one agent. Distributed optimization problems were first studied systematically in [1] where the union of the graphs was assumed to be strongly connected among each time interval of a certain bounded length and the adjacency matrices were doubly stochastic. A distributed subgradient method was introduced to solve the distributed optimization and error bounds on the performance index functions were given. As a continuation of [1], a distributed subgradient projection algorithm was developed in [2] for distributed optimization where each agent was constrained to remain in a closed convex set and the paper gave corresponding convergence analysis on identical closed convex sets and on fully connected graphs with non-identical closed convex sets. Inspired by the works of [1, 2], the algorithms proposed in [1] and [2] were studied in the random environment [3] and [4], where the agents had the same state constraint. In [5], the communication topology was undirected and each possible communication link was functioning with a given probability. Thus, the expected communication topology is essentially fixed and undirected. Different from [1]-[5], a dual averaging subgradient algorithm was developed and analyzed for randomized graphs under the assumption that all agents remain in the same closed convex set in [6] and it was shown that the number of iterations were required by their algorithm scales inversely in the spectral gap of the network. Moreover, distributed optimization problems with asynchronous step-sizes or inequality-equality constraints or using other algorithms were studied in [7]-[12] and corresponding conditions were given to ensure the system converge to the optimal point or its neighborhood. However, as in [1]-[5], it was assumed in [6]-[12] that the state sets of agents to be identical or the objective function finally converge to only a neighborhood of the optimal set.

In this paper our work is to extend [14] to study the penalty primal-dual subgradient projection algorithm in a more general method. In [14], the authors solved a multi-agent convex optimization problem where the agents subject to a global inequality constraint, a global equality constrain and a global constraint set. In order to solved these constraints, the author in [14] presented two different distributed projection algorithms with three assumptions that the union of the graphs is assumed to be strongly connected among each time interval of a certain bounded length and the adjacency matrices were doubly stochastic and non-degeneracy. However, [14] guaranteed the edge weight matrices of graphs were doubly stochastic (i.e.,  for all  and , and  for all  and ). Previous work did not perform well on the application of the distributed algorithms in multi-agent network.

Contributions: The subgradient algorithm (we proposed) is different with the approach proposed in [14] in properties and analysis. In our approach, the communication topology is without loss of generality. This paper does not recur to the assumption that the adjacency matrices are doubly stochastic, and we only require the network is weight-balanced, which makes our algorithm more practical. In this paper, we consider a general multi-agent optimization problem where the main focus is to minimize a global objective function which is a sum of local objective functions, subject to global constraints, including an inequality constraint, an equality constraint and a (state) constraint set. Each local objective function is convex and only known by one particular agent. On the other hand, the inequality (resp. equality) constraint is given by a convex (resp. affine) function and known by all agents. Each node has its own convex constraint set, and the global constraint set is defined as their intersection. Particularly, we assume that the local constraint sets are identical. Our main interest is in computing approximate saddle points of the Lagrangian function of a convex constrained optimization problem. To set the stage, we first study the computation of approximate saddle points (as opposed to asymptotically exact solutions) by using the subgradient method with a constant step-size. We consider constant step-size rule because of its simplicity and practical relevance, and because our interest is in generating approximate solutions in finite number of iterations.

The paper is organized as follows. In Section II, we give some basic preliminaries and concepts. Then, in Section III, we present our problem formulation as well as distributed consensus algorithm preliminaries. We then introduce the distributed penalty primal-dual subgradient algorithm with some supporting lemmas and continue with a convergence analysis of the algorithm in Section IV. Furthermore, the properties of the algorithm are explored by employing a numerical example in Section V. Finally, we conclude the paper with a discussion in Section VI.

2. Preliminaries and Notations

In this section, we first introduce some preliminary results about graph theory, the properties of the projection operation on a closed convex set and convex analysis (referring to [13], [14]).

A.     Algebraic Graph Theory

The communication among different nodes in an information interplay network can be modeled as a weighted directed graph , where  is the set of nodes with  representing the th node,  is the edge set, and  is the weighted adjacency matrix of  with nonnegative adjacency elements  and zero diagonal elements. A directed edge  implies that node  can reach node  or node  can receive information from node . If an edge , then node  is called a neighbor of node  and . The neighbor node set of node  is denoted by , while we define  as the number of neighbors of node . The Laplacian matrix  associated with the adjacency matrix  is defined by ; , which ensures that . The Laplacian matrix  has a zero eigenvalue, and the corresponding eigenvector is . Note that the Laplacian matrix  of a directed graph  is asymmetric. The in-degree and out-degree of node  can be respectively defined by the Laplacian matrix as :  and . A directed path from node  to node  is a sequence of edges  in the directed graph  with distinct nodes . A directed graph is strongly connected if for any two distinct nodes  and  in the set , there always exists a directed path from node  to node . A graph is called an in-degrees (or out-degrees) balanced graph if the in-degrees (or out-degrees) of all nodes in the directed graph are equal. A directed graph with  nodes is called a directed tree if it contains  edges and there exists a root node with directed paths to every other node. A directed spanning tree of a directed graph is a directed tree that contains all the network nodes.

B.     Basic Notations and Concepts

The following notion of saddle point plays a critical role in our paper.

Definition 1 (Saddle point): Consider a convex-concave function , where ,  and  are closed convex subsets in  and . We are interested in computing a saddle point  of  over the set , where a saddle point is defined as a vector pair  that satisfies

, for all

In this paper, we do not assume the function  at some points are not differentiable, and the subgradient plays the role of the gradient.

Definition 2 : For a given convex function  and a point , a subgradient of the function  at  is a vector  such that the following subgradient inequality holds for any :

Similarly, for a given concave function  and a point , a supgradient of the function  at  is a vector  such that the following supgradient inequality holds for any :

We use  to denote the projection of a vector  on a closed convex set , i.e.

In the subsequent development, the properties of the projection operation on a closed convex set play an important role. In particular, we use the projection inequality, i.e., for any vector

 for all        (1)

We also use the standard non-expansiveness property, i.e.

 for any  and        (2)

In addition, we use the properties given in the following lemma.

Lemma 2.1: Let  be a nonempty closed convex set in . Then, we have for any ,

(a)    , for all .

(b)    , for all .


(a)    Let  be arbitrary. Then, for any , we have

By the projection inequality [cf. (1)], it follows that


, for all

(b)    For an arbitrary  and for all , we have

By using the inequality of part (a), we obtain

, for all

Part (b) of the preceding lemma establishes a relation between the projection error vector and the feasible directions of the convex set  at the projection vector.

The following notations besides those aforementioned will be used throughout this paper.  denotes the set of all -dimensional real vector spaces. Given a set , we denote  by its convex hull. We write  or  to denote the transpose of a vector  or a matrix . We let the function   denote the projection operator onto the non-negative orthant in . Denote  and . For a vector , we denote , while  is the standard Euclidean norm in the Euclidean space. In this paper, the quantities (e.g., functions, scalars and sets) associated with agent  will be indexed by the superscript .

3. Problem Statement

We consider a multi-agent network model. The nodes connectively at time  can be represented by a directed weighted graph , where  is the set of activated edges at time , i.e., edge  if and only if node  can receive data from node , and  is the adjacency matrix, in which is the weight assigned to the edge  at time . Please note that the set  is the set of edges with non-zero weights . In this paper the agents are to correspondingly solve the following optimization problem:


where  is a convex objective function of agent , and  is a nonempty, closed, compact and convex subset of . In particular, we study the cases where the local constraint sets are identical i.e.,  for each agent, and  is a global decision vector. Assume that  is only known by agent . The function  is known by all the agents with each component , for , being convex. The inequality  is component-wise; i.e., , for all , and represents a global inequality constraint. The function , represents a global equality constraint, and is known by all the agents. Let  denote the optimal value of (3) and  denote an optimal solution of (3). We assume that the optimal value  to be finite. We also represent the optimal solution set by , i.e., . We will assume that in general  is non-differentiable.

To generate optional solutions to the primal problem of Eq. (3), we consider optional solutions to its dual problem. Here, the dual problem is the one arising from penalty relaxation of the inequality constraints  and equality constraints . Note that the primal problem (3) is trivially equivalent to the following:

with associated dual problem given by

Here  is the penalty dual function defined by , where  is the penalty function given by . We often refer to vector  with  as two multiplier. We denote the dual optimal value by  and the dual optimal set by . We define the penalty function  for each agent  as follows:  . In this way, we have that . We say that there is zero duality gap if the optimal value of the primal and the dual problems are equal, i.e., . As proven in the following lemma, the Slater’s condition in Assumption 3.1 ensures zero duality and the existence of penalty dual optimal solutions.

Assumption 3.1 (Slater’s Condition): There exists a vector  such that  and . And there exists at least one interior  of , i.e. , problem (3) has finite optimal solution, and  has nonempty interior point.

Lemma 3.1: Let the Slater condition holds, the values of  and coincide, and  is non-empty.

Proof: Define Lagrangian function  as , with the associated dual problem defined by


Here, the dual function, . The dual optimal value of problem (7) is denote by  and the set of dual optimal solutions is denoted by . Since  is convex,  and , for , are convex, and  is finite and the Slater’s condition holds, we can conclude that  and . We now proceed to characterize  and . Pick any . Since , then


On the other hand, pick any . Then  is feasible, i.e.,   and . It implies that  holds for any  and , and thus . Therefore, we have .

To prove the non-empty of , we pick any . From (5) and , we can see that  and thus .

Throughout this paper, we use the following assumption for problem (3).

Assumption 3.2: Let the following conditions hold:

 The set  is closed and convex.

 Each function  is convex.

 All functions  have Lipschitz gradients with a constantfor all .

 The gradients  are bounded over the set , i.e., and there exists a constant  such that  for all  and all .

When each  has Lipschitz gradient with a constant , assumption 3.2(3) is satisfied with . When  is compact, the Assumption 3.2(4) holds. We here make the following assumptions on the network communication graphs .

Assumption 3.3 (Non-degeneracy): There exists a constant  such that , and , for , satisfies , for all .

Assumption 3.4 (Weight-balanced):  is weight-balanced if , for all .

Assumption 3.5 (Periodical Strong Connectivity): There is a positive integer  such that, for all , the directed graph  is strongly connected.

Lemma 3.2 Saddle-point Theorem: The pair of  is a saddle point of the function  over  if and only if it is a pair of primal and penalty dual optimal solutions and the following penalty minimax equality holds:

Based on this characterization, we will use the subgradient method of the following section for finding the saddle points of the penalty function. We denote , for each  and we define the function  as . Note that  is convex in  by using the fact that a nonnegative weighted sum of convex functions is convex. For each , we define the function  as . It is easy to check that  is concave in . Then the penalty function  is the sum of convex-concave local functions.

Lemma 3.3 (Dynamic Average Consensus Algorithm) [21] : The following is a vector version of the first-order dynamic average consensus algorithm with :

We set  for . The sequences of  satisfy  and . Suppose that periodical strong connectivity Assumption 3.5 holds. Assume that  for all  and all . Then  for all .

Proof: Define

where  is referred to as the reference signal (or input) of node  at time .

We propose the First-Order Dynamic Average Consensus Algorithm below to reach the dynamic average consensus:

Let  and  be fixed. Then for every , there exists a real number  such that for every integer , and , it holds that for



Without loss of generality, we only consider the case where , being identical with the proof for a general . Fixing some , it holds that

Let , we have that


Since (8) holds for all ,


Applying recursive method, it follows that


Since at every , we have that


where we are using the property of (10) in the last two inequalities. Applying repeatedly (11), we have that, for any integer , the following holds for

where .

Now we proceed by induction on . Suppose that (6) holds for some ; then we should show (6) for . By the induction hypothesis, we have that for all integer , there exists some  such that the following holds for

Consequently, as in (11), we have

Following along the same lines as in (11), we obtain  for all  where  and . This establishes (6) for . By induction, we have shown that (6) holds. The proof for (7) is analogous.

Let , then  for any . By replacing  and  in (4) with  and  respectively. We have that for every

Similarly, we can see that

Combining the above two inequalities gives that

Denoting  for an integer . From (9), we know that . Thus we have

where .

For any , let  be the largest integer such that , and . Thus for all  it follows that


Since  and  are input-to-output stable with ultimate bound ; i.e., there exist  and such that

Choosing as initial state  for all . Since , we can deduce that


It follows from (13) that  and thus

Let , for any . The implementation of the Dynamic Average Consensus Algorithms ensures that . So we can conclude that

Thus,  holds.

Consider the following Distributed projected subgradient algorithm proposed in [13]: Suppose  is a closed and convex set. Let . Denote . The following is a slight modification of Lemma 8 and its proof in [13].

Lemma 3.4: Let the non-degeneracy Assumption 3.3, the weighted-balanced Assumption 3.4, and the periodic strong connectivity Assumption 3.5 hold. Then there exist  and  such that

Suppose  is uniformly bounded for each , and , then we have .

4. Distributed Subgradient Methods

In this section, we introduce a distributed penalty primal-dual subgradient algorithm to solve the optimization problem (3), followed by its convergence properties.

Distributed Penalty Primal-Dual Subgradient Algorithm

We consider a set  of agents. Each agent chooses any initial state , and . At any time , each agent  computes the following convex combination:

and updates its estimates , and  according to the following ways:


where the scalars  are nonnegative weights and the positive scalars  are step-sizes,  is the projector onto the set . The vector  is a subgradient of the agent  ‘s penalty function  at , where  is the convex combination of dual estimates of agent  and its neighbors’.  keeps to the following rules:

Remark 4.1: Since , it follows that

where  is the Laplacian matrix such that .

Proof: Modifying the second term on the right-hand side in the above formula, we then have

Let , one has

Since graph  is balanced, then  and . We can conclude that  and  hold under the condition that  satisfies . Similarly, we obtain ,  and .

Assumption 4.1 (Step-size assumption): The step-sizes satisfy , ,  and , , .

In the following, we study the convergence behavior of the subgradient algorithm introduced in this section where the optimal solution and the optimal value is asymptotically agreed upon.

Theorem 4.1 (Convergence properties of the DPPDS algorithm): Consider the problem (3). Let the non-degeneracy Assumption 3.3, the weight-balanced Assumption 3.4 and the periodic strong connectivity Assumption 3.5 hold. Consider the sequences of  and  of the distributed penalty primal-dual subgradient algorithm, where the step-sizes  satisfy the step-size Assumption 4.1. Then there exists a primal optimal solution  such that  for all . Furthermore, we have  for all .

Remark 4.2: The distributed penalty primal-dual subgradient algorithm takes the equality constraint into account. The presence of the equality constrain can make  unbounded. Therefore, unlike other subgradient algorithm, e.g., [15], [16], the distributed penalty primal-dual subgradient algorithm does not involve the dual projection steps onto compact sets. So we do not guarantee the subgradient  not to be absolutely bounded, while the boundedness of subgradients is a standard assumption in the analysis of subgradient methods, e.g., see [6], [13], [17], [18], [19], [20]. The step-size of Assumption 3.1 is stronger than the more standard diminishing step-size scheme in [22] and this will correctly deal with the difficulty of the boundedness of . We give this condition in order to prove, in the absence of the boundedness of , the existence of a number of limits and summability of expansion toward Theorem 4.1. Finally, we adopt the penalty relaxation instead of the Lagrangian relaxation in this paper.

Remark 4.3 (Penalty subgradient inequality): Observe that  and  (due to the fact that is convex and ). Moreover,  is a supgradient of ; i.e. the following penalty supgradient inequality holds for any  and :


Proof: Observe that  holds for all , ,  and  is arbitrarily. Thus,


Followed by the properties of supgradient, we obtain

Remark 4.4: In this paper, we apply the harmonic series  into our subgradient algorithm. Its easy to check that  satisfies the step-size Assumption 4.1 (for more details, one may refer to [14]).

A         Convergence Analysis

In the following, we will prove convergence of the distributed penalty primal-dual subgradient algorithm. First, we rewrite our algorithm into the following form:

where  is projection error described by

and , ,  are some local inputs. Denote the maximum deviations of dual estimates by  and . We further denote the averages of primal and dual estimates as  and . Since  is compact, and  and  are continuous, there exist  such that for all , it holds that  for all ,  and . Since  is a compact set and , ,  are convex, then it follows from Proposition 5.4.2 in [19] that there exist  such that for all , we have that ,  and .

Lemma 4.1: Let . Consider the sequence  defined by  , where ,  and .

(a)          If , then .

(b)          If , then .

The proof of Lemma 4.1 can be referred to Lemma 5.1 in [14].

Lemma 4.2 (Diminishing and summable properties): Suppose the weighted-balanced Assumption 3.4 and the step-size Assumption 4.1 hold.

(a) It holds that , and the sequences of ,  and  are summable.

(b) The sequences of , , ,  and  are summable.

Proof: (a) Noticing that

Then, we show that

Recalling that ,. This implies that the following inequalities hold for all :

Then we deduce the following recursive estimate on . Repeatedly applying the above estimates yields that


where .

Similar arguments can be employed to show that


Since  and , then we know that  and . Noticing that

Then, the following estimate on  holds:


Recalling that ,  and . Then we have . By (16), we obtain

It follows from the step-size Assumption 4.1 that . Similarly, one can show that . Multiplying both sides of  by  and square, then we deduce the following recursive estimate:

Then the summability of ,  and  testifies that of .

(b) Noticing that


Then following from Lemma 3.4 with  and , we have the summability of . Then  is summable. Similarly, it holds that .

We now consider the evolution of . Recalling that . By Lemma 2.1, let ,  and , we get

Regrouping the above estimates, we obtain

With the above relation, from Lemma 3.4 with  and , the following holds for some  and :


Multiplying both side of (20) by  and using (18), for all , it yields

By applying the relation of  and sorting out, we get

Part (a) gives that  is summable. Meanwhile,  are bounded, and , then we can say that the first term on the right-hand side in the above estimate is summable. Recalling that , it’s easy to check that the second term is also summable. Part (a) gives that  and  is summable. Then according to the Lemma 7 in [13] with , ensure that the third term is summable. In summary,  is summable. Following the same lines in (19), one can show the summability of . Similarly,  and  are summable.

Lemma 4.3 (Basic iteration relation): The following estimates hold for any  and :




Proof: By Lemma 2.1, we can deduce that . Let , we have

Expanding and regrouping the above formula, we obtain

Owing to the subgradient inequality ,

it follows that:

Lemma 4.4 (Achieving consensus): Let assumption 3.3-3.5 holds. Consider the sequences of  and  of the distributed penalty primal-dual subgradient algorithm with the step-size sequence  and the associated  satisfy , . Then there exists  such that  for all . Furthermore, , and  for all .

Proof: By Lemma 4.3, we see that

Owing to , one can show that

Since , taking the limit on  in the above inequality, then it follows that:

and thus  exists for any :

On the other hand, taking limits on both sides of the above inequality, we have  and therefore we deduce that  for all . It follows from Lemma 4.1 that  for all . Combining this with the property, we can deduce that  for all .

Since , is continuous, for ,  and , we can deduce that , , .

Claim 1: For any  and , the sequences of  and are summable

Proof: Observing that

Recalling that , we then have

By using the summability of  and  in part (b) of Lemma 4.2, we have that  are summable. Similarly, the following estimates hold:

Due to ,  and  holds for all , the following estimates hold:

Then the property of  in part (b) in Lemma 4.2 implies the summability of the sequence  and that of


Claim 2: Denote the weighted version of the local penalty function  over  as . The following property holds: .

Proof: Summing (21) over  and replacing  by , we can deduce that


The summability of  in Part (b) implies that the right-hand side of (21) is finite as , and thus


On the other hand,  is a saddle point of  over . Since, then we have . Claim 1 and (24) renders that

and thus . On the other hand,  (due to  is convex) implies . Similarly, we have the following estimates . Thus .

Claim 3: Denote . And we denote the weighted version of the global penalty function  over  as

The following property holds: .

Proof: Noticing that


By using the boundedness of subgradients and the primal estimates, we can see that

Noticing that ,  and  holds for all , the following estimates hold:


Then it follows from (b) in Lemma 4.2 that  is summable. Notice that . Following the Claim 2, hence, .

Claim 4: The limit point  in Lemma 4.4 is a primal optimal solution.

Proof: Let . By  and , we get


This indicates that the sequence  is non-decreasing in . Observing that  is lower bounded by zero. Therefore, we give the following two cases:

Case 1: The sequence  is upper bounded. Then  is convergent in . Then it follows from Lemma 4.4 that  for all . This implies that there exists  such that  for all . Recalling that  in (27). Following a recursive step, we can get . Since and , we obtain . Since  for all , we have  for all , then  and thus .

Case 2: The sequence  is not upper bounded. Since  is non-decreasing, then  by . Recalling that  and , then it follows from (a) in Lemma 4.1 that it is impossible that . Suppose that . Then we obtain


Taking limits on both sides of (28), then we get

Finally, we reacha contradiction, implying that .

In both cases, we obtain  for any . Similarly, we can further prove . Since , then  is feasible and thus . For another, since  is a convex combination of  and , then it follows from Claim 3 and (b) in Lemma 4.1 that:

Hence, we have  and thus .

Lemma 4.5: It holds that .

Proof: Since , , then the following holds for any

The following can be proven by induction on  for a fixed :


Let  in (29) and recall that initial state  for all . Then we obtain


From (30), we can obtain


Combining (31) with  gives the desired result. Based on the above five Lemmas, we then finish the prove of Theorem 4.1.

5. Numerical Example

In this section, we study a simple numerical example to illustrate the effectiveness of the proposed distributed penalty primal-dual subgradient algorithm. Consider a network with five agents. Suppose each agent  has a function , given by

where the global decision vector . The global inequality constraint function is given by , the global equality constraint function is give by  and the global constraint set is given as: .  are parameters of , whose values are randomly choosen from the intervals . Consider the optimization problem as follows:


We solve problem (32) by employing the distributed penalty primal-dual subgradient algorithm (14) with the step-size . Its simulation results are shown from Figs. 1 to 5. It can be seen from Fig. 1 that local input  tends to  when it achieves consensus. Fig. 2 shows the state evolutions of all five agents, which demonstrate that all agents’ takes  iterates to asymptotically achieve consensus. The state evolutions of dual solution  and  are shown in Figs. 3 and 4, respectively. We can observe from Fig. 5 that all the agents asymptotically achieve the optimal value.

Fig. 1. Local input  tends to  when achieve consensus.

Fig. 2. Optimal solution  of primal problem.

Fig. 3. Optimal solution  of dual problem.

Fig. 4. Optimal solution  of dual problem.

Fig. 5. Optimal solution  of objective function .

6. Conclusion and Future Work

In this paper, we formulated a distributed optimization problem with local objective functions, a global equality, a global inequality and a global constraint set defined as the intersection of local constraint sets. In particular, we considered the local constraint sets to be identical. Then, we proposed a distributed penalty primal-dual subgradient algorithm for the constrained optimization with a convergence analysis. Moreover, we employed a numerical example to show that the algorithm was asymptotically converge to primal solutions and optimal values. Future workmay aim at the analysis that the local constraint sets of each agent are imparities. Also, we will pay attention to the convergence rates of the algorithms in this paper.


This work described in this paper was supported in part by the Natural Science Foundation Project of Chongqing CSTC under grant cstc2014jcyjA40041, in part by the Scientific and Technological Research Program of Chongqing Municipal Education Commission under KJ1501401.


  1. J.C.Duchi, A.Agarwal,andM.J.Wainwright. Dual averaging for distributed optimization: convergence analysis and network scaling,IEEE Transactions on Automatic control, 2012, 57(3): 592-606.
  2. A.Nedić, A.Ozdaglar,andP.Parrilo. Constrained consensus and optimization in multi-agent networks,IEEE Transactions on Automatic Control, 2010, 55(4): 922-938.
  3. K.SrivastavaandA.Nedić. Distributed asynchronous constrained stochastic optimization,IEEE Journal of Selected Topics in Signal Processing, 2011, 5(4): 772-790.
  4. S. S. Ram, A.Nedić and V. V. Veeravalli, "Distributed Stochastic Subgradient Projection Algorithms for Convex Optimization," Journal of optimization theory and applications, 2010, 147(3): 516-545
  5. S.S.Ram, A.Nedić,andV.V.Veeravalli. Distributed stochastic subgradient projection algorithms for convex optimization,Journal of optimization theory and applications, 2010, 147(3): 516-545.
  6. A.NedićandA. Ozdaglar. "Distributed Subgradient Methods for Multiagent Optimization", IEEE Transactions on Automatic Control, 54(1): 48-61,2009.
  7. S.S.Ram, A.Nedić,andV.V.Veeravalli. Incremental stochastic subgradient algorithms for convex optimization,SIAM Journal on Optimization, 2009, 20(2): 691-717.
  8. J.Lu, C.Y.Tang, P.R.Regier,andet al. A gossip algorithm for convex consensus optimization over networks,American Control Conference (ACC), 2010. IEEE, 2010: 301-308.
  9. M. Zhu and S.Martínez."An approximate dual subgradient algorithm for distributed non-convex constrained optimization", in the proc. of Conference on Decision and Control, 2010, pp. 7487-7492.
  10. M.RabiandM.Johansson. A simple peer-to-peer algorithm for distributed optimization in sensor networks,IEEE Conference onDecision and Control. 2007, 46: 4705-4710.
  11. E.Wei, A.Ozdaglar,andA.Jadbabaie. A distributed newton method for network utility maximization,IEEE Conference onDecision and Control (CDC). 2010, 49th: 1816-1821.
  12. A.Nedić. Asynchronous broadcast-based convex optimization over a network,IEEE Transactions on Automatic Control, 2011, 56(6): 1337-1351.
  13. A.Nedić, A.Ozdaglar,andP.Parrilo. Constrained consensus and optimization in multi-agent networks,IEEE Transactions on Automatic Control, 2010, 55(4): 922-938.
  14. M.ZhuandS.Martínez. On distributed convex optimization under inequality and equality constraints via primal-dual subgradient methods,arXiv preprint arXiv:1001.2612, 2010.
  15. B.Johansson, T.Keviczky, M.Johansson,and K. H. Johansson. Subgradient methods and consensus algorithms for solving convex optimization problems, InIEEE Conference onDecision and Control,pages 4185-4190, Cancun, Mexico, December2008.
  16. A.Nedić, A.Olshevsky, A.Ozdaglar,andet al. Distributed subgradient methods and quantization effects,IEEE Conference on Decision and Control. 2008: 4177-4184.
  17. A.NedićandA.Ozdaglar. Subgradient methods for saddle-point problems,Journal of optimization theory and applications, 2009, 142(1): 205-228.
  18. A.NedićandA.Ozdaglar. Approximate primal solutions and rate analysis for dual subgradient methods,SIAM Journal on Optimization, 2009, 19(4): 1757-1780.
  19. D.P.Bertsekas. Convex optimization theory,Athena Scientific, 2009.
  20. D.P. Bertsekas, A.Nedić,andA.Ozdaglar. Convex analysis and optimization,Athena Scientific,2003.
  21. M.Zhu,andS.Martínez. Discrete-time dynamic average consensus,Automatica, 2010, 46(2): 322-329.
  22. A.NedićandA.Ozdaglar. Subgradient methods in network resource allocation: Rate analysis,Information Sciences and Systems, IEEE Conference on Decision and Control,2008: 1189-1194.

Article Tools
Follow on us
Science Publishing Group
NEW YORK, NY 10018
Tel: (001)347-688-8931