Distributed Constrained Optimization Towards Effective Agent-Based Microgrid Energy Resource Management

Abstract

The current energy scenario requires actions towards the reduction of energy consumption and the use of renewable resources. In this context, a microgrid is a self-sustained network that can operate connected to the smart grid or in isolation. The long-term scheduling of on/off cycles of devices is a critical problem that has been commonly addressed by centralized approaches. In this work, we propose a novel agent-based method to solve the long-term scheduling problem as a distributed constraint optimization problem (DCOP) by modelling future system configurations rather than reacting to changes. Moreover, with respect to approaches based on decentralised reinforcement learning, we can directly encode system-wide hard constraints (such as for example the Kirchhoff law) which are not easy to represent in a factored representation of the problem. We compare different multi-agent DCOP algorithms showing that the proposed method can find optimal/near-optimal solutions for a specific case study.

Applicability of Multi-Agent Systems and Constrained Reasoning for Sensor-Based Distributed Scenarios: A Systematic Mapping Study on Dynamic DCOPs

“…The problem to solve is how to perform a continuous distributed orchestration and/or a load balancing of the involved services, where dynamics are represented as variations over time in the demand for such services. Smart Grids [ 23 , 61 , 62 ] scenarios, where agents manage and control different elements from the electricity grid such as generators, dispatchable loads, feeders, etc., and dynamics are represented as variations in the energy supply and demand in consecutive time windows. Fioretto et al define the Distributed-EDDR problem to model the Economic Dispatch and the Demand Response as two parallel sub-problems being merged in order to solve it through a DPOP-like algorithm solver [ 61 ].…”

“…Fioretto et al define the Distributed-EDDR problem to model the Economic Dispatch and the Demand Response as two parallel sub-problems being merged in order to solve it through a DPOP-like algorithm solver [ 61 ]. Lezama et al’s approach was based on the resolution of sequential canonical DCOPs with policies that enable agents to maintain or update their domains and constraints between consecutive time windows [ 23 , 62 ]. Traffic flow [ 63 , 64 , 65 ] consists of performing plans and scheduling systems over transiting vehicles in order to avoid jams and/or bottlenecks.…”

“…Given this extension, they designed both complete and incomplete algorithms based on DPOP and S-DPOP predecessors [ 25 , 54 ]. Framework [ 17 , 43 , 55 , 62 ]: We identified a set of papers that propose diverse structures or problem definitions to handle Dynamic DCOPs, where a off-the-shelf classic DCOP is assembled instead a pure novel algorithm. The DCOP_MST model introduces changes in constraints through the inclusion of dynamic domains [ 17 ].…”

“…DCEE problems were defined as a framework to introduce and coordinate two algorithms for exploration and exploitation ofsensing agents [ 55 ]. Lezama et al proposed a framework where classic DCOPs on consecutive time windows are connected by an inter-phase operation where a set of helper agents updates information from dynamics in parallel with the current state of DCOP agents [ 62 ].…”

“…The most recent proposals (from 2015 to 2020) are the most appropriate to adopt today as this set has been applied in various IoT- and sensor-based scenarios [ 17 , 46 , 56 , 57 , 58 , 59 , 62 ] and include the Stability and Anytime mechanisms spread over different algorithms [ 42 , 47 , 56 , 62 , 68 ] that provide robustness on environments with a high frequency of dynamics.…”