Combinatorial optimization in Biology using Probability Collectives Multi-agent Systems

We present a study of Probability Collectives Multi-agent Systems (PCMAS) for combinational optimization problems in Biology. This framework for distributed optimization is deeply connected with both game theory and statistical physics. In contrast to traditional biologically-inspired algorithms, Probability-Collectives (PC) based methods do not update populations of solutions; instead, they update an explicitly parameterized probability distribution p over the space of solutions by a collective of agents. That updating of p arises as the optimization of a functional of p. The functional is chosen so that any p that optimizes it should be p peaked about good solutions. In this paper we demonstrate PCMAS as a promising combinational optimization method for biological network construction. This computational approach to response networks enables robust prediction of activated crucial sub-networks in biological systems under the presence of specific drugs, thereby facilitating the identification of important nodes for potential drug targets and furthering hypotheses about biological and medical problems on a systems level. The application of PCMAS in this context therefore sheds light on how this multi-agent learning methodology advances the current state of research in agent-based models for combinational optimization problems in Biology.

► Show the effectiveness of advanced agent-based machine learning algorithms (PCMAS).
► PCMAS can solve important combinatorial optimization problems in Biology.
► Drug targets may be identified using biological networks constructed by PCMAS.
► PCMAS facilitates hypotheses about biological/medical problems on a systems level.
► Mycobacterium tuberculosis is the case study for PCMAS to show its optimization strength.