Dynamic Pricing based Efficient Model for Smart Parking System

Abstract

The concept of Smart Parking (SP) has emerged as a beacon of hope for innumerable vehicles navigating congested streets in today’s urban environment. Utilizing cutting-edge technology, SP alleviates the persistent difficulties associated with locating and securing a parking location. Despite its promise, this innovative solution confronts a number of challenging obstacles. Increasing urbanization and vehicle ownership have exacerbated the parking space shortage, making efficient Parking Management (PMGMT) more important than ever. Inadequate infrastructure, data privacy concerns, and the integration of various technologies are obstacles that require careful consideration. In the domain of SP solutions, establishing a balance between motorist convenience and sustainable urban development remains a formidable challenge. The on-street parking system and Electric Vehicle (EV) charging facilities at public places pose more challenges than any other form. This work examines the literature on Dynamic Fare Pricing (DFP), Dynamic Charging/ Discharging Pricing (DCDP) for EVs, Dynamic Parking Pricing (DPP), and Dynamic Congestion Pricing (DCP) based resource allocation in the field of Intelligent Transportation Systems (ITS). The presented review contributed in numerous ways by detailing the benefits and drawbacks of allocation techniques, the evolution of techniques, various evaluation parameters, taxonomies, strengths and weaknesses, tools, data-sets, and a comparative analysis of allocation techniques. Here, two dynamic Pricing and Allocation Schemes (PASs), OccPARK and DyPARK and one dynamic Congestion control and Allocation Scheme (CAS) CoPARK, are developed. In all schemes, immediate and advance parking or parking-cum-charging requests are generated, and decisions on them are made. Locally, at a single Parking Controller (PC) or Charging Controller (CC), all requests are processed in the order they were received. The decision on advance requests can be delayed based on requester’s preferences and availability of resources. The OccPARK determines dynamic parking prices based on the actual and predicted occupancy of requested Parking Lot (PL). The benchmark prices are determined using parking prices data set of the Seattle city. In this scheme, the dynamic parking prices increases or decreases with the increase or decrease in occupancy levels respectively. The DyPARK scheme takes into account two categories of Parking Users (PUs): Paid Parking Users (PPUs) and Restricted Parking Users (RPUs). The later ones have free access to the parking spaces at least once per day. This scheme solves four objectives, i.e., minimization of parking prices for PPUs, maximization of free parking duration for RPUs, maximization of generated revenue for Parking Authorities (PAs), and minimization of the parking duration loss for PAs. In addition, the algorithms are designed to generate dynamic prices for PPUs and free parking durations for RPUs. This scheme is fair because it provides the Nash Equilibrium (NE) solution, which allows all parties to pursue their objectives. The CoPARK scheme addresses the issue of charging allocation for EVs at limited public Charging Stations (CSs). It allocates Charging Ports (CPs) and the amount of charge to be given using Deep Reinforcement Learning (DRL), specifically Deep Deterministic Policy Gradient (DDPG). In contrast, the technique for Multi-Criteria Decision Analysis (MCDA), Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), is used to reduce a multi-objective problem to a single objective one. While allocating CPs to EVs, this scheme optimizes a number of objectives, including the minimization of charging prices, the maximization of generated revenue, the minimization of charging load on CPs, the maximization of CP availability, the minimization of waiting time at CPs, and the minimization of Service Level Agreement (SLA) violations or penalty costs. Using equitable Dynamic Pricing (DP) methods, charging prices and penalty costs are calculated. The OccPARK and DyPARK schemes are evaluated using a simulation environment of Seattle on-street parking system. In contrast, the CoPARK scheme is evaluated by considering the road network of Sioux Falls under the assumption that CSs are present at each node. The DyPARK and CoPARK schemes are also implemented using cloud, fog, and edge computing paradigms. As a result, distinct communication protocols were devised to facilitate communication between the various components of the parking and EVs charging system. The worth of the proposed OccPARK, DyPARK, and CoPARK schemes was determined using various evaluation parameters by comparing them to the finest available state-of-the-art schemes in the literature.