University of Twente Student Theses


Improving logistic planning in the construction industry using Wireless Sensor Networks and Multi Agent Systems

Engels, Dirk (2010) Improving logistic planning in the construction industry using Wireless Sensor Networks and Multi Agent Systems.

Abstract:Construction sites are heavily dependent on logistics. Also many construction processes are highly dependent on each other and the supply of building resources. A delay in the beginning of the supply chain could lead to a lack of supplies at the end which could cause a halt. Planning the logistics of the construction industry is a challenging task. Large savings can be made by increasing planning efficiency. This research focuses on the logistic planning of Remix which produces dry mortar for the construction industry. Remix places silos at construction sites and refills them when needed. Currently the silo statuses are checked manually and on irregular intervals. This results in high priority orders, which are more difficult to plan on time with a tight schedule. In this research the supply chain of the construction industry has been analyzed in order to improve the logistic planning. Wireless Sensor Networks are used for measuring the environment. Wireless sensor networks can support the delivery company managing orders by providing information about inventory levels. In this case WSN are used for gathering silo information located at construction sites. This information supports the planning department to prevent silo stock outs. Multi Agent Systems can also use this data to create a view of the environment in which they need to act. Depending on their level of control they can alert the foreman when a silo gets to a critical level or placing an order automatically. A prototype application has been developed to show the benefit of using information gathered by WSN. Planners can easily view critical silo statuses and contact the construction site for placing an order on time. Also foremen of construction sites can use the prototype to view their silo statuses and place orders when needed. Such an application reduces the time needed for entering orders, because orders can be placed by the customers themselves. Also the availability of the information will lead to fewer urgent orders. Also a simulation application has been built to simulate the usage of intelligent agents and evaluate its performance. A set of hypotheses are tested against the output of several simulation runs with each different input settings representing different scenarios. The result of this simulation provides an impression of the performance of a multi agent system for the logistic planning of the construction industry. The prototype application clearly shows the usefulness of data gathered by the wireless sensor network and providing this information to the planners. Planners can now actively view silo statuses and take action when needed. Applying Multi Agent Systems for the logistic planning is possible and can result in performance improvements. However in order to make such a system an improvement of the current situation the agents have to deal with a lot of information and make the right decision with this information. The simulation tool developed in this research simulates the processes in the supply chain of a mortar company. Different scenarios have been analyzed and evaluated with different settings for the agents. The outcome provides a better understanding of the forces influencing the performance of the agents. Implementing such a solution in the business environment requires a few changes to the planning process. Depending on the level of control of the agents, agents can place orders automatically 4 without interference of the planner. The role of the planner changes from planning actively to verify the planning made by the agents. This reduces the time needed to create an optimal planning, which can cope well with changes in an existing planning. An optimal planning also reduces fuel costs and the amount of man power needed.
Item Type:Essay (Master)
Faculty:BMS: Behavioural, Management and Social Sciences
Subject:85 business administration, organizational science
Programme:Business Information Technology MSc (60025)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page