Overall Optimization of Smart City by Multi-population Global-best Brain Storm Optimization using Cooperative Coevolution

This paper proposes a method for the overall optimization of smart city (SC). The proposed method is based on multi-population global-best brain storm optimization using cooperative coevolution (MP-CCGBSO). Using a SC model, energy cost, actual power loads during peak periods, and carbon dioxide emission can be minimized. For the SC problem, many researchers have proposed various evolutionary algorithms including CCGBSO, which applied cooperative coevolution to GBSO. However, there is still room to improve quality of the solution by CCGBSO. Taking Toyama city of Japan as the research object, the calculation results of original CCGBSO method and the proposed MP-CCGBSO method of 2, 4, 8 and 16 populations are compared.

I. Introduction

Global warming has caused many disasters in recent years, and climate emergency is worsening every day. The emergency includes extreme high temperature, air pollution, wildfires, intensified floods, and drought, and it is affecting lives of people all over the world. Reasons of global warming include the excessive emission of greenhouse gases [1]. Therefore, it is necessary to advance renewable energies in order to reduce carbon dioxide emissions and traditional fossil energies. Many countries are implementing and validating SC demonstration projects to reduce carbon dioxide emissions [2] [3]. SC is a sustainable low carbon city using renewable energy, batteries, and the latest information technology. The climate change conference (COP25), held in Madrid, Spain, in 2019, considered ways to further reduce carbon dioxide emissions and strengthen the implementation of the Paris Agreement [4]. Since it is difficult to evaluate the actual carbon dioxide emission reduction and energy cost in the development of SC, it is necessary to establish a model to evaluate it. Industry, building, residence, and railway sectors are modeled by two models respectively. One is a dynamic model considering the transient phenomena in various sectors. Another one is a static model considering all kinds of energy balance. However, there is no SC model that can solve the calculation of carbon dioxide emissions or energy consumption in all sectors at the same time. Therefore, experts in various sectors have developed SC models to quantitatively assess the energy cost or carbon dioxide emissions of the whole SC. However, considering the interaction between various sectors, the optimization of SC energy network had not been applied to these models.