(A) minimal cost design and performance analysis for automated parking system

Abstract

Recently, conventional parking systems seldom meet a growing number of parking demands, especially in the urban areas which have limited parking space. Automated parking systems, which automatically park and retrieve vehicles, have been steadily replacing conventional parking systems. As compared to conventional parking systems, the automated parking systems have many advantages such as efficient utilization of parking space, improved vehicle safety, decrease in vehicle damages, and better control of vehicles. On the other hand, they are very expensive to build and inflexible to future change and so a very careful design of the system is justified. Among various types of automated parking systems, Rotary Parking System (RPS), Elevator Parking System (BPS) and Shuttle Parking System (SPS) are selected for the study due to their popularity in real world. This thesis deals with design and performance analysis of the three systems. The first part of the thesis develops the expected service time models. Using the steady-state probabilities obtained from the continuous time Markov chain, the expected service times for RPS and EPS are developed under Closest Open Location (COL) policy. Additionally, we develop an efficient aggregation method for obtaining the probability that storage bay is full using Markov chain for EPS. For SPS, analytical expressions of the expected service time are derived by continuous representation under randomized storage assignment policy. To examine the performance of the proposed models, comparative studies are performed. The second part addresses the design problem of automated parking systems. The economic design models are developed in non-linear integer programming with the objective of minimizing the total cost. However, the models are incapable of representing dynamic nature of the parking system, especially the associated levels of a given set of performance measures. Therefore, simulation models are run from which the required perf...