Multi-agent collaborative search with Tchebycheff decomposition and monotonic basin hopping steps

This paper presents a novel formulation of Multi Agent Collaborative
Search for multiobjective optimization. A population of agents combines
global exploration heuristics and moves to explore the neighborhood of
each agent. In this novel formulation the selection process is based on
the Tchebycheff decomposition of the multiobjective optimization problem
into single objective optimization problems in combination with the
use of the dominance index. The decomposition allows the implementation
of Monotonic Basin Hopping steps that improve convergence on
single funnel structures. The novel agent-based algorithm is tested on
a standard benchmark and on a real space trajectory design problem.
Its performance is compared against a number of state-of-the-art multiobjective optimization algorithms.