2006 Graduate Catalog and 2005 Annual R & D Report - Sirindhorn ...

2006 Graduate Catalog and 2005 Annual R & D Report
Sirindhorn International Institute of Technology (SIIT)
Research Interests:
GA-based Design of Manual Assembly Lines with
Respect to Productivity and Postural-Physical
Loads Smoothness
Traditionally, an assembly line is designed based on
a single factor, namely, productivity. More
specifically, the design problem involves the
determination of the number of assembly
workstations and the assignment of assembly tasks to
individual workstations based on the given line
throughput such that the balance delay is minimized.
The worker-workstation assignments are normally
fixed and the assembly tasks that workers have to
perform are repetitive. For a manual assembly line,
such requirements put workers at high risk of
cumulative trauma disorders (CTDs) due to
unbalanced postural and physical workloads on the
musculoskeletal system of the body. This research
project is intended to apply a genetic algorithm (GA)
approach to the design of a manual assembly line
such that system productivity and postural-physical
loads smoothness are concurrently optimized. The
Rapid Upper Limb Assessment (RULA) technique is
applied to ergonomically evaluate the postural and
physical loads imposed by individual assembly tasks.
The project seeks to explain the chromosome
encoding, evaluation procedure, crossover, mutation,
and selection procedure to generate new generations
that are superior to the previous ones. The results
from the GA-based design procedure will yield the
number of workstations and the worker-workstation
assignments for the assembly line.
Performance Analysis of Ergonomics-based
Manual Assembly Line with Parallel Workstations
and Floaters
It has been long known that manual assembly tasks
are repetitive and require the use of specific muscles
in the upper extremities, creating excessive postural
and physical loads on the excessively used body
members. As a result, manual assembly line workers
are at high risk of cumulative trauma disorders in the
upper extremities. When ergonomics concerns are
introduced to the manual assembly line balancing
problems to assign assembly tasks to workers to
achieve the maximum postural-physical loads
smoothness, the resulting task-workstation
assignment solution must be determined using an
ergonomics-based approach. Unfortunately, the line
throughput is likely to decrease since the balance
delay of the line might be compromised. This
research project is intended to investigate the effect
of adding parallel workstations to some potential
bottleneck workstations and providing floaters (or
extra helpers) to those parallel workstations so as to
increase the assembly line productivity. Additionally,
several strategies for opening parallel workstations
and rotating floaters among them will be investigated
with respect to various desired throughput rates. The
Rapid Upper Limb Assessment (RULA) technique will
be used to assess the postural and physical loads
imposed on the musculoskeletal system of the body
when performing each manual assembly task. A
heuristic procedure will be employed to assign
assembly tasks to workstations using a mixed
productivity and ergonomics consideration. Based on
predefined dispatching strategies to assign floaters
and rotate them among parallel workstations,
simulation models will be developed. The
performance of the given manual assembly line under
different throughput rate requirements and
operational conditions will be analyzed. The variables
of interest include: throughput rate, number of fulltime
workstations, number of parallel workstations,
number of floaters, dispatching strategy, and
switchover strategy. The performance indices are:
achieved throughput rate, balance delay, utilization
rates of full-time and parallel workstations, switchover
rates, and average queue lengths at full-time
workstations.
Ergonomics-based Workforce Scheduling for the
Vehicle Routing Problem
The vehicle routing problem is intended to determine
the optimal number of vehicles to delivery goods
between finite sets of origins and destinations, and
their delivery routes. There are numerous variants of
the vehicle scheduling problem that have been
studied by operations research and industrial
engineering researchers. However, very few (if any)
have paid attention to the vehicle drivers. In realworld
situations, vehicle drivers might not only drive
delivery vehicles but also perform loading and
unloading of goods at both the origins and
destinations. With limited time windows,
loading/unloading operations may require more than
one person to perform. Moreover, long-distance
driving is stressful and increases the risk of highway
accidents. Alternate drivers may be required for
certain delivery routes. This research project is
intended to take the loading/unloading workload and
long-distance driving into consideration when finding
the optimal workforce schedule for the vehicle routing
problem. Based on the given delivery loads (in terms
of required energy costs) and the driving distances for
individual delivery trucks, a heuristic approach will be
developed to determine the minimum numbers of
vehicles and operators (drivers and movers) and their
delivery routes so as to minimize the total traveling
distance without exceeding the recommended daily
energy expenditure and driving distance.
Decision Support System for Selecting the Noise
Hazard Control Strategy
This research project is intended to develop a
decision support system (DSS) for selecting the noise
hazard control strategy for any noisy workplace when
the budget is fixed. The DSS program will consist of
the database module, input module, computing
module, and output module. The database module
will contain noise source data, noise data, feasible
noise control methods and their implementation costs,
and worker locations. The input module will enable
the user to enter the total noise control budget,
allocated budget portions to the engineering approach
and to the use of hearing protection devices (HPDs),
and preferred noise control methods. The computing
module will utilize the genetic algorithm (GA) and
heuristic approaches to optimally select the
30