On Distributed Stochastic Gradient Algorithms for Global Optimization
Autor: | Anirudh Sridhar, Brian Swenson, H. Vincent Poor |
---|---|
Jazyk: | angličtina |
Rok vydání: | 2019 |
Předmět: |
FOS: Computer and information sciences
0209 industrial biotechnology Optimization problem Computer science Computation MathematicsofComputing_NUMERICALANALYSIS 02 engineering and technology 010501 environmental sciences 01 natural sciences Maxima and minima Set (abstract data type) symbols.namesake 020901 industrial engineering & automation Optimization and Control (math.OC) Gaussian noise FOS: Mathematics symbols Key (cryptography) Computer Science - Multiagent Systems Global optimization Algorithm Mathematics - Optimization and Control Multiagent Systems (cs.MA) 0105 earth and related environmental sciences |
Zdroj: | ICASSP |
Popis: | The paper considers the problem of network-based computation of global minima in smooth nonconvex optimization problems. It is known that distributed gradient-descent-type algorithms can achieve convergence to the set of global minima by adding slowly decaying Gaussian noise in order to escape local minima. However, the technical assumptions under which convergence is known to occur can be restrictive in practice. In particular, in known convergence results, the local objective functions possessed by agents are required to satisfy a highly restrictive bounded-gradient-dissimilarity condition. The paper demonstrates convergence to the set of global minima while relaxing this key assumption. |
Databáze: | OpenAIRE |
Externí odkaz: |