A decentralized multi-objective optimization algorithm

Abstract: During the past two decades, multi-agent optimization problems have drawn increased attention from the research community. When multiple objective functions are present among agents, many works optimize the sum of these objective functions. However, this formulation implies a decision regarding the relative importance of each objective function. In fact, optimizing the sum is a special case of a multi-objective problem in which all objectives are prioritized equally. In this paper, a distributed optimization algorithm that explores Pareto optimal solutions for non-homogeneously weighted sums of objective functions is proposed. This exploration is performed through a new rule based on agents' priorities that generates edge weights in agents' communication graph. These weights determine how agents update their decision variables with information received from other agents in the network. Agents initially disagree on the priorities of the objective functions though they are driven to agree upon them as they optimize. As a result, agents still reach a common solution. The network-level weight matrix is (non-doubly) stochastic, which contrasts with many works on the subject in which it is doubly-stochastic. New theoretical analyses are therefore developed to ensure convergence of the proposed algorithm. This paper provides a gradient-based optimization algorithm, proof of convergence to solutions, and convergence rates of the proposed algorithm. It is shown that agents' initial priorities influence the convergence rate of the proposed algorithm and that these initial choices affect its long-run behavior. Numerical results performed with different numbers of agents illustrate the performance and efficiency of the proposed algorithm.

Review text

In this article is proposed a new gradient-based optimization algorithm and all necessary arguments, such as proof of convergence to solutions and convergence rates are presented. The algorithm is for multi-agent multi-objective set constrained problems. The agents have an initial vector of priorities (weights), and an initial vector of decision variables. The weights determine how the agents update the decision variables according with information received from the other agents. The proposed algorithm surveys the Pareto front and performs four steps at each iteration as follows:\begin{itemize} \item[1)] agent i updates its vector of priorities using those received from the other agents in the network, \item[2)] the vectors of priorities generate the matrix of information weights for the decision variable update, \item[3)] agent i updates its vector of decision variables with variables received from its neighbors, \item[4)] agent i takes a gradient descent step and projects its new decision variables onto the constraint set.\end{itemize} All necessary mathematical arguments are presented in this work or in the annex. The effectiveness of the proposed algorithm with different agents is illustrated by some numerical results, presented in the last part of the paper.

Recommendations

Cites work

Cited in

(8)

This page was built for publication: A decentralized multi-objective optimization algorithm

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2032002)