[USArmy2019] Multitasking Evolutionary Algorithms for Optimizing Artificial Neural Network and Graph-Based Models
Information
- PI: Assoc. Prof. Huynh Thi Thanh Binh
- Time: 2019-2022
- Funding: U.S Army Research Lab and US Army International Research Center – Asia Pacific
- Main areas: Optimization, Graph, Evolutionary Computing
Intro
A wide range of practical applications in economics, industry, science and other fields have been modeled as combinatorial optimization problems on graphs such as routing problems, construction of irrigation systems, product distribution, network designing problems, and so on.
Two major approaches are usually applied in order to deal with these combinatorial optimization problems, which are exact algorithms and approximation algorithms. Recently, the Multifactorial Evolutionary Algorithm (MFEA) has been emerging as an effective approximation algorithm to deal with various fields of problems. The MFEA, which is a variant of Evolutionary Algorithm (EA), can solve multiple independent optimization problems simultaneously using a single representation of population.
In this project, fundamental researches are applied to solve some combinatorial optimization problems on graph namely Clustered Shortest-Path Tree Problem, Clustered Minimum Routing Cost Problem, Inter-Domain Path Computation under Domain Uniqueness Constraint, Data aggregation in Wireless sensor networks, Charging Scheme for Energy Depletion Avoidance in WRSNs, etc. Targeting these problems, we will mainly focus on studying various advances of the MFEA, which helps to handle multiple problems at the same time.
