Yanchun Pan

Optimizing green production strategies: An integrated approach

Selection of green production strategy is a critical but difficult task due to the fact that it affects not only green benefits, but also production economy. The problem is essentially multi-objective and involves dynamic and uncertain conditions. This study focused on an integrated approach to improve the analysis and facilitate decision-making process. Discrete-event simulation model was developed to capture production flow and decision logic under real world conditions. A multi-objective genetic algorithm (MOGA), combined with improving heuristics, was developed to search the best solutions (Pareto optimums). The two modules are integrated to work in evolutionary cycles to achieve the optimization. Experiments were designed and carried out via a prototype system developed to verify and validate proposed concepts, including sensitivity analysis of related model parameters.

Simulation based analysis for selection and evaluation of green manufacturing strategies

Evaluation and selection of green sustainable development (GSD) strategies is a critical but difficult process involving many uncertain factors and affected by dynamic conditions. The decision problem is essentially multi-objective. This research proposed a simulation based method that uses simulation to capture the complicated process flow and decision logic involved, and interface the simulation with a search procedure to improve solutions. The design of simulation model emphasized a robust structure that represents adequately the process flow and decision logic of typical applications of GSD improvement. A process/activity based cost structure was built into the simulation to capture essential trade-offs contributed by different cost drivers during the process of GSD implementation. Experiments were designed and implemented via simulation to verify and validate proposed concepts. An analysis of response surface method (RSM) was conducted to illustrate the search for optimal solutions under relaxed conditions, which showed the potential of developed research.

Enterprise behaviour under Cap-and-Trade conditions: an experimental study with system dynamic models

Operating under Cap-and-Trade programme conditions has brought new challenges to industrial organizations pursuing green sustainable development by imposing more constraints on resource/energy acquisition and disposition in order to reduce green-house-gas emissions. New factors due to the conditions interact with each other and system production/service functions, causing complicated dynamic relationships that significantly affect the overall performance of participating enterprises. Managers now have to balance production economy and green improvement. This research shows a proactive approach based on system dynamics modelling to represent such complex systems and analyse the relationship to explore the underlining logic that drives system behaviour under related conditions. The analysis provides useful insights for decision or policymakers to pursue environmentally friendly and economically sound production. Conceptual models were developed and verified for functions, and simulation experiments were designed to implement the models and compare different strategies to analyse their impact on the system’s overall performance, such as long-term over-emission, continuous green investment on emissions reduction, and cost of purchasing emissions allowance or paying penalty (via over-emission tax).

A simulation optimization framework for shipment planning at RDC considering time and quantity consolidation with uncertain demands

Shipment planning (SP) at regional distribution center (RDC) involves order consolidation and vehicle routing decisions under uncertain demands, which is generally very hard to be solved by traditional analytical methods such as mathematical programs. To cope with the complexity of this important problem existing in logistics systems, a general-purpose simulation optimization framework is proposed. Discrete-event simulation (DES) is employed to model the complicated shipping processes and capture the systems dynamics and uncertainties. A new policy (ID-policy) considering time and quantity consolidation is developed to improve consolidation effectiveness. The consolidated orders and systems performance obtained by simulation are then transformed as input into a genetic algorithm designed to optimize the vehicle routes via evolutionary computation. Experiment results show that the ID-policy outperforms traditional consolidation policies such as T-policy, Q-policy and D-policy under different conditions. The proposed simulation optimization framework is also validated by the exemplary case.

An ARIMA based model for forecasting the patient number of epidemic disease

Forecasting the number of epidemic disease is very important for CDC (center for disease control and prevention). To improve the forecast accuracy, an ARIMA (autoregressive integrated moving average) based model is proposed in this paper. First, autocorrelation (AC) and partial autocorrelation (PAC) analysis are introduced to establish a stationary time series, where the autocorrelation order, moving average order and difference order are estimated. Secondly, least square s method (LS) is employed to estimate the parameters of the prediction model. Finally, the real data between Jan. and Aug. 2014 coming from a CDC are fed into the proposed model and the forecast accuracy obtained is 92.1%, which significantly outperforms the simple moving average method currently used in the CDC.

Green production - strategies and dynamics: a simulation based study

There are a number of issues for enterprises to implement green production. From operations perspective, selecting green improvement strategy is critical but difficult due to the fact that it affects not only green performance, but also production economy. Important trade-off exists between different objectives and decisions are subjected to dynamic and uncertain conditions. From system dynamics perspective, there exist multiple factors interacting with one another to drive systems behavior and the trade-offs. Decision makers need to evaluate different scenarios to find appropriate balance between strategies. We report studies addressing both issues through an integrated approach emphasizing the use of simulation. First it focused on the optimization of green improvement strategies. A simulation model was developed to capture operations flow and decision logic. A multi-objective genetic algorithm, combined with improving heuristics, was developed to search for best solutions. Secondly, system dynamic models were developed to characterize the dynamic behavior of production systems under Cap and Trade conditions. Simulation experiments were run to analyze the relationship between system states and among the factors that cause the state transitions that influence the overall system behavior.

Multi-agent based simulation of carbon emissions trading market in China

With the development of international carbon emissions trading market, the establishment of carbon trading market is an inevitable trend in China. This research aims to analyze a hypothesized carbon emissions trading market by using multi-agent based system simulation. Multi-agent based modeling simulates the behaviors of interacting and autonomous agents and their impact. In our study, the software-Anylogic© is used to build a model which contained 100 agents (enterprises) interacting with one another in an emission-quote trading market as autonomous decision makers. These agents act as buyers or sellers. A market-control agent directs the trading between the buyer and seller agents and enables a market mechanism with dynamic auction.

Capacitated multi-modal network flow models for minimizing total operational cost and CO2e emission

Abstract Green logistics and transportation impose new requirements on distribution planning and flow optimization over complex logistics network. Carbon-dioxide equivalent (CO2e) emission on a road, for instance, depends on the transportation mode and vehicle type selected along the road. The handling cost of a load at a transshipment city also depends on the change of transportation mode and vehicle type from inbound to outbound at the city. Compared with traditional network structure, multi-modal network is characterized by multi-attributes link and dependent transshipment node (in terms of minimizing total cost and emission), which significantly increases problem complexity. This study attempts to address key issues in modeling and analysis of such systems. Analytical models are developed to characterize the problem structure/features explicitly and integrate decisions at single- and multi-objective levels. A set of experiments, designed based on real transportation logistics network, are carried out to verify and validate models’ functions, analyze sensitivity of parameters, and evaluate solution convergence behavior. The results indicated that while the problem is NP-hard, solutions via proper analytical models can be derived efficiently for problems of practical scale.

A rule-based system to support carbon resource planning under C&T conditions

As part of a larger research project, this study addresses issues of risk-based decision-making based on different decision-makers in carbon resource planning for manufacturers under the constraints of C&T, we specify their characterization of behavior, attitude to the C&T program, and design their rule-based expert system (RBS) respectively. we develop discrete-event simulation models for conservative decision-maker and proactive decision-maker to conduct experimental analyzes. More than three years of efforts have been made on extensive literature review and empirical studies (such as site-visits and meetings) to identify the problem and define system models for intended analysis. A baseline model has been implemented with ARENA© and limited experiments were conducted to verify the basic functions. While the preliminary results confirmed with the expectations about the system behavior and validated the original ideas, tremendous work remains in terms of model implementation, modification and experiments.

Multimodal transportation network optimization with environmental and economic performance considered: An ongoing research

An optimization model is proposed in this paper to solve the transportation mode selection and the demand allocation problem with environmental and economic performance considered. The case study based on the regional logistics of Pearl River Delta (PRD) in Guangdong province of China illustrates that the optimal solution changes significantly when carbon emission is introduced into the optimization objective. More ocean and less ordinary road transportation will be adopted when environmental performance matters, and it is possible to greatly reduce the carbon emission while keeping the transportation cost not increase dramatically. However, there exist a tradeoff between emission reduction and the transportation cost. The proposed model can also be extended to take the transportation time and due dates of demands into consideration in the process of multimodal transportation network optimization.