Peak load minimization of an e-bus depot: impacts of user-set conditions in optimization algorithms

Energy Informatics

Table 1 Overview of Existing Literature (Continued)

Research	Type of Power Load	Objective	Method	Preemption	V2G	Time Granularity (min)	Offline/Real-Time	Real Data	Fleet Size
(Gao et al. 2018)	Bus	Cost	GA	Y	N	15	Off	Y	581
(Yang et al. 2019)	Car	Cost	Heu (Search&Swap)	Hybrid	N	15	RT	N	1000
(Koutsopoulos and Tassiulas 2012)	Generic power load	Cost	Optimal Control	N	N	[5,60]	RT	N	20
(Xie et al. 2016)	EV + House	Peak	MDP+Adaptive DP	Y	Y	12	RT	N	1000
(Škugor and Deur 2015)	Truck	Cost	DP	N	N	60	Off	Y	10
(Mhaisen et al. 2020)	Car	Cost	RL	Y	Y	60	RT	N	1
(Wang et al. 2019)	Bus	Cost	MDP	Y	N	-	RT	Y	16359
This Work	Bus	Peak	MILP	Y/N	Y/N	[15,30,60]	Off	Y	138

ILP — Integer Linear Programming; LP — Linear Programming; QP — Quadratic Programming; MPC — Model Predictive Control; DP — Dynamic Programming; Heu — Heuristic; GA — Genetic Algorithm; MDP — Markov Decision Process; RL — Reinforcement Learning; MILP — Mixed-Integer Linear Programming