Submodular optimization is a special class of combinatorial optimization arising in several machine learning problems, but also in cooperative control of complex systems. In this paper, we consider agents in an asynchronous, unreliable and time-varying directed network that aim at cooperatively solving submodular minimization problems in a fully dis- tributed way. The challenge is that the (submodular) objective set-function is only partially known by agents, that is, each one is able to evaluate the function only for subsets including itself. We propose a distributed algorithm based on a proper linear programming reformulation of the combinatorial problem. Our algorithm builds on a column generation approach in which each agent maintains a local candidate basis and locally generates columns with a suitable greedy inner routine. A key interesting feature of the proposed algorithm is that the pricing problem, which involves an exponential number of constraints, is solved by the agents through a polynomial time greedy algorithm. We prove that the proposed distributed algorithm converges in finite time to an optimal solution of the submodular minimization problem and we corroborate the theoretical results by performing numerical computations on instances of the s–t minimum graph cut problem

Distributed Submodular Minimization Over Networks: A Greedy Column Generation Approach / Testa, Andrea; Notarnicola, Ivano; Notarstefano, Giuseppe. - ELETTRONICO. - (2018), pp. 4945-4950. (Intervento presentato al convegno IEEE Conference on Decision and Control tenutosi a Miami Beach, Florida, USA nel 17-19 December 2018) [10.1109/CDC.2018.8618958].

Distributed Submodular Minimization Over Networks: A Greedy Column Generation Approach

Testa, Andrea
;
Notarnicola, Ivano;Notarstefano, Giuseppe
2018

Abstract

Submodular optimization is a special class of combinatorial optimization arising in several machine learning problems, but also in cooperative control of complex systems. In this paper, we consider agents in an asynchronous, unreliable and time-varying directed network that aim at cooperatively solving submodular minimization problems in a fully dis- tributed way. The challenge is that the (submodular) objective set-function is only partially known by agents, that is, each one is able to evaluate the function only for subsets including itself. We propose a distributed algorithm based on a proper linear programming reformulation of the combinatorial problem. Our algorithm builds on a column generation approach in which each agent maintains a local candidate basis and locally generates columns with a suitable greedy inner routine. A key interesting feature of the proposed algorithm is that the pricing problem, which involves an exponential number of constraints, is solved by the agents through a polynomial time greedy algorithm. We prove that the proposed distributed algorithm converges in finite time to an optimal solution of the submodular minimization problem and we corroborate the theoretical results by performing numerical computations on instances of the s–t minimum graph cut problem
2018
2018 IEEE Conference on Decision and Control
4945
4950
Distributed Submodular Minimization Over Networks: A Greedy Column Generation Approach / Testa, Andrea; Notarnicola, Ivano; Notarstefano, Giuseppe. - ELETTRONICO. - (2018), pp. 4945-4950. (Intervento presentato al convegno IEEE Conference on Decision and Control tenutosi a Miami Beach, Florida, USA nel 17-19 December 2018) [10.1109/CDC.2018.8618958].
Testa, Andrea; Notarnicola, Ivano; Notarstefano, Giuseppe
File in questo prodotto:
File Dimensione Formato  
submodular_columngen_post_reviewed.pdf

Open Access dal 22/07/2019

Tipo: Postprint
Licenza: Licenza per accesso libero gratuito
Dimensione 690.67 kB
Formato Adobe PDF
690.67 kB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11585/674736
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? 2
social impact