Using Systems Dynamics Formalism as Base for an Innovative Hybrid Modeling Approach: Methodology and Case Study

摘要

Simulation is the best tool used for any non-trivial, real world system. For analysis of complex systems, simulation is often used prior to the operation of the real world system as a mediator for a dynamic situation. Therefore, simulation methodology has been recommended and chosen to analyze, for example, container terminal systems. This paper considers and compares two models for port simulation approach: Discrete-Event Simulation (DES) and System Dynamics (SD). Both simulation approaches started and evolved almost simultaneously with the advent of computers, but very little communication existed between these fields. This is however, changing at present with more DES or SD academics and practitioners showing an interest to enter the other world. DES models systems as a network of queues and activities, where state changes occur at discrete points of time, whereas SD models consist of a system of stocks and flows where continuous state changes occur over time. In DES the objects (entities) are individually represented and can be tracked through the system. Specific attributes are assigned to each entity and determine what happens to them throughout the simulation. On the other hand, in SD entities are presented as a continuous quantity. In DES state changes occur at discrete points of time, while in SD state changes happen continuously at small segments of time (At). Specific entities cannot be followed throughout the system. DES models are stochastic in nature with randomness incorporated through the use of statistical distributions. SD models are generally deterministic and variables usually represent average values. Despite the differences listed, it is claimed that the objective of models in both simulation approaches is to understand how systems behave over time and to compare their performance under different conditions. Existing work on the comparison of DES and SD is scarce. In the few studies found, comparisons tend to be biased towards either the DES or SD approach. The views expressed consist mainly of the authors' personal opinions based on their own area of expertise. Little understanding exists regarding the differences and similarities between the two simulation approaches, let alone understanding when should one approach be used instead of the other. The ports are very intensive production realities respect to regular facilities: 24h work time per day, 365 days working in a year, all-weather operations are just some of the stressing factor of this sector; port facilities involve big investments and require special operative and management skills. Both models (SD and DES) are used in this study to resolve problem of port simulation (in Arabian gulf region). The goal is to provide a study about two different models that are not alternative but can be integrated. The models considered in this study are two. The goal of Dry Bulk model consists into evaluate the Dry Bulk storage capacity for every kind of material in order to ensure the total annual throughput, in according with the feasibility study assumption reported into the Master plan document about berths, dock cranes and transportation means. The goal of Container terminal model consists into evaluate the container stacking/storage capacity best configuration in order to ensure the total annual throughput and minimize the shuffling moves, standing the feasibility study assumption reported into the Master plan document about berths, dock cranes and transportation means. Modular approach has been considered resulting in a 10-stocking yard for the first stage configuration (up to 2028-2030) and 20 stocking yard for the final configuration (over 2038). The model developed considers, for each stocking yard, the following modules: shipping module, stocking yard module, horizontal transportation module (with Terminal Tractor). For each stage configuration the evaluated output are: number of container and TEUs (Twenty-Foot Equivalent Unit)/ year (total annual throughput), number of container and TEUs / ship (average parcel size), ship calls / year, average number of STS cranes moves / hour, average yard utilization coefficient, average container crossing time, average stacking height (for each storage area block served by one RTG). The simulation run length considered is 2 years for each scenario. The results of this study suggest that it is possible use both approaches: Discrete-Event Simulation and System Dynamics to simulate the port.