WSC 2007 Final Abstracts

Simulation-based Scheduling Track

Monday 10:30:00 AM 12:00:00 PM
Applied Simulation-Based Scheduling

Chair: Anders Skoogh (Chalmers University of Technology)

A Simulation-based Framework for Quantifying the Cold Regions Weather Impacts on Construction Schedules
Adham Shahin, Simaan AbouRizk, and Yasser Mohamed (University of Alberta) and Siri Fernando (The City of Edmonton, Drainage Services)

In cold regions, weather severity has a major impact on construction activities carried out in the open, leading to significant deviations from baseline schedules. Faced with weather uncertainty, setting up the project baseline schedule, against which performance will be measured, can be challenging. Construction planners often depend on their personal judgment and experience to account for the scheduling impact of cold weather. Due to variation in planners' experiences, the regions where their experience was acquired, and the time of year when the planned project is to be executed, the result can be widely differing plans between planners. This paper presents a structured simulation-based approach that attempts to account for the cold weather impacts on project schedule. This approach relies on generating weather sequences similar to the existing historical weather data for the project's geographical region. The impact on productivity and project schedule can then be modeled, leading to consistent schedules.

Simulation Assisted Match-up Rescheduling of Flexible Production Systems Subject to Execution Exceptions
Wilhelm Dangelmaier, Kiran R. Mahajan, and Mark Aufenanger (University of Paderborn) and Thomas Seeger (Accenture)

An immense amount of research work has been done in the areas of scheduling and re-scheduling of various types of manufacturing systems. In this paper we present a simulation assisted approach to rescheduling complex production system configurations subject to execution exceptions. Issues like how to bring the deviation of a schedule due to exceptions back to its original trajectory and how to do this in real-time without affecting coordination problems on the shop floor are addressed. Our results show that combining simulation and optimization for rescheduling indeed helps to achieve both these objectives and that this approach proves to be promising to help reduce chaos in today's dynamic manufacturing environments.

Reflective Simulation for On-line Workload Planning and Control
Roberto Revetria and Flavio Tonelli (DIPTEM University of Genoa)

Since its beginning, simulation has been used to study complex systems in order to infer on their future behavior, in this field several applications have been made using it as off-line tool for strategic level choices. In modern application, especially in the field of industrial automation, on-line simulation has been extensively used for supporting operative decision trough a classical schedule-simulate loop. The paper presents an application of on-line simulation to the distribution logistics sector: a department store is here controlled by on-line simulators able to help decision maker to decide how many counters to kept opened or how many people to use for shelves replenishment. Since this exercise could seriously affect the performances of a real life department store, the methodology is, in fact, very sensible to parameter settings, a nested simulator has been implemented and used for algorithm fine tuning and critical parameter choice.

Monday 1:30:00 PM 3:00:00 PM
Simulation-Based Scheduling Algorithms

Chair: Edward Williams (University of Michigan Dearborn)

Stochastic Rollout and Justification to Solve the Resource-Constrained Project Scheduling Problem
Ningxiong Xu and Linda Nozick (Cornell University) and Orr Bernstein and Dean Jones (Sandia National Labs)

The key question addressed by the resource-constrained project scheduling problem (RCPSP) is to determine the start times for each activity such that precedence and resource constraints are satisfied while achieving some objective. Priority rule-based heuristics are widely used for large problems and more recently justification has been shown to be an important extension. Xu et al. further augments priority rule heuristics by creating rollout procedures and proves their effectiveness. However, that procedure generates just one schedule. We extend that method using sampling to generate a set of schedules using probabilistic techniques and select the best schedule from this sample. Using the 600 problem instances in PSLIB, we present empirical evidence that this procedure produces solutions that are better than the rollout procedure alone but at a computational cost.

Online Multiobjective Single Machine Dynamic Scheduling with Sequence-dependent Setups Using Simulation-based Genetic Algorithm with Desirability Function
Adeline T. H. Ang and Appa Iyer Sivakumar (Nanyang Technological University)

This paper presents a Simulation-based Genetic Algorithm with Desirability function (SIMGAD) that could be used on-line for the dynamic scheduling of a single machine with sequence-dependent setups. The weights used to combine the criteria (dispatching rules) into a single rule using linear weighted aggregation is determined by genetic algorithm (GA). The GA evaluates the performance of each set of weights with discrete-event simulation that returns a fitness value after multiple performance measures (objectives) are each expressed as a desirability function and combined into a single objective function. An illustrative simulation example based on the scheduling of an ion implanter machine in wafer fabrication plant shows that SIMGAD works effectively in solving the multiobjective scheduling problem with capability of handling user preference in decision making to achieve the desired performances.

A Metaheuristic Algorithm for Simultaneous Simulation Optimization and Applications to Traveling Salesman and Job Shop Scheduling with Due Dates
George Jiri Mejtsky (Simulation Research)

We describe a metaheuristic algorithm for simulation optimization. Traditionally, discrete event simulation optimization is carried out by multiple simulation runs executed sequentially. At the end of each simulation run, the run is evaluated (using model output, black box approach) by an objective function. If we carry out simulation runs simultaneously, then we can evaluate (using model internal data, white box approach) different simulation runs during their execution before the end is reached. Thus, we can eliminate the inferior runs early and allow only the most promising runs to continue to the end. We explore this parallel competition of simulation models on a single processor computer. Applications of the algorithm to traveling salesman and job shop scheduling problems are presented. In conclusion, our results suggest that the algorithm is a suitable approach for solving some combinatorial problems, and it represents a promising nonsequential avenue for simulation optimization.

Monday 3:30:00 PM 5:00:00 PM
User Friendliness

Chair: Toni Ruohonen (University of Jyvaskyla)

A Web-based Simulation Optimization System for Industrial Scheduling
Marcus Andersson, Henrik Grimm, Anna Persson, and Amos Ng (University of Skovde)

Many real-world production systems are complex in nature and it is a real challenge to find an efficient scheduling method that satisfies the production requirements as well as utilizes the resources efficiently. Tools like discrete event simulation (DES) are very useful for modeling these systems and can be used to test and compare different schedules before dispatching the best schedules to the targeted systems. DES alone, however, cannot be used to find the "optimal" schedule. Simulation-based optimization (SO) can be used to search for optimal schedules efficiently without too much user intervention. Observing that long computing time may prohibit the interest in using SO for industrial scheduling, various techniques to speed up the SO process have to be explored. This paper presents a case study that shows the use of a Web-based parallel and distributed SO platform to support the operations scheduling of a machining line in an automotive factory.

Modeling and Simulation for Customer Driven Manufacturing System Design and Operations Planning
Juhani Heilala and Jari Montonen (VTT Technical Research Centre of Finland), Arttu Salmela (Raute Corporation) and Pasi Jarvenpaa (Sandvik Mining and Construction Finland Oy)

Agility, speed and flexibility in production networks are required in today's global competition in the flat world. The accuracy of order date delivery promises is a key element in customer satisfaction. Agile production needs a management and evaluation tool for production changes. Discrete event simulation, DES, has mainly been used as a production system analysis tool, to evaluate new production system concepts, layout and control logic. DES can be used for operational planning as well, as shown in the paper. The simulation analysis gives a forecast of the future with given input values, thus production managers have time to react to potential problems and evaluate alternatives. A balance between multiple parallel customer orders and finite resources can be found. The authors are developing a system design evaluation method and also a decision support system for production managers. Two case studies with different approaches are described in the paper.

Simulation Improves End-of-line Sortation and Material Handling Pickup Scheduling at Appliance Manufacturer
Edward J. Williams (PMC), Onur M. Ulgen (University of Michigan - Dearborn) and Marcelo C. Zottolo and Neelesh Kale (PMC)

The most venerable and the most highly varied general application area in which simulation has frequently and repeatedly proved its economic value is the manufacturing sector of the economy. Manufacturing applications of simulation have included attention to complex issues of equipment and/or worker downtime, problems of facility layout, work and line balancing, bottleneck analysis, and material handling. Furthermore, simulation has proved itself capable of addressing productivity and efficiency improvement tasks in which these complexities overlap and interact. Historically, much of the success simulation has enjoyed in other economic sectors (e.g., service, transportation, and health care) has stemmed in large measure from the reputation it earned in the manufacturing sector. The manufacturing application described in this paper proved the cost-effective feasibility of designing sortation operations downstream of an assembly line, and scheduling SKU pickups there, with no risk of blockage of that line.

[ Return to Top ]