Abstract
The main objectives of this research are investigating the process planning knowledge of allocating construction resources for earthmoving processes and capturing this knowledge in an Intelligent Resource Allocation System (IRAS). The specific focuses of this research are capturing planning knowledge in selecting an appropriate equipment class, capturing planning experiences of allocating construction resources, and automating simulation studies. IRAS consists of three modules to accomplish the focuses of this research. These modules are an expert system, a pattern classification, and a Goal Driven Simulation System (GDSS). The functions of the expert system module are to capture planning knowledge in selecting a construction technology (e.g. scrapers) and an equipment class (e.g. conventional class of scrapers) and to recommend an appropriate equipment class based on process data (e.g. travel distance). The functions of the pattern classification modules are to capture the experiences of allocating construction resource, to recommend a preliminary resource assignment, and to adapt new experiences. An Artificial Neural Network (ANN) using Backpropagation algorithm is used as the pattern classification tool. A justification for selecting ANN and Backpropagation algorithm is presented. The functions of the GDSS module are to allocate a final resource assignment and to automate simulation studies. GDSS consists of an integration of a simulation program and an analysis program that analyzes a simulation output and synthesizes an appropriate resource assignment. GDSS reduces the user's knowledge requirements in conducting simulation studies to specifying the goal (e.g. moving 300 CY/HR), the model data, and the number of equipment units. Using the goal as the ultimate objective, GDSS conducts a simulation study through alternating between the simulation program and the analysis program until the goal is satisfied and the assigned equipment are effectively utilized. The analysis and simulation programs are developed using C_language. The modules of IRAS prototype have been validated using an earthmoving process. GDSS module is restricted to automate the execution and analysis phases of a simulation study. Research results demonstrated that IRAS prototype shows a viable first phase toward fully automating the resource allocation process.
Alsugair, Abdullah Mohammed (1992). An intelligent resource allocation system. Texas A&M University. Texas A&M University. Libraries. Available electronically from
https : / /hdl .handle .net /1969 .1 /DISSERTATIONS -1281164.