Research Summary

UNDERGRADUATE THESIS (B.TECH PROJECT)
OBJECT ORIENTED FRAME WORK FOR SCHEDULING USING NON TRADITIONAL OPTIMIZATION TECHNIQUES
Guide : Dr. Arbind Kumar Singh, IIT Guwahati.

Brief Summary
CPM and PERT techniques assume unlimited availability of construction resources which are limited in practice. Activities can be scheduled to comply with constrained resource usage, after the critical path has been identified. The other aim is to make the resource usage as smooth as possible. Some times different time estimates are available for activities. Although ower project duration leads to higher direct costs, indirect costs are also decreased. In such a case it is important to study the trade off between completion time and cost of the project. These problems, when formulated mathematically as optimization problems, are $\mathcal{NP} - Hard$. To overcome the complexity, these problems have been treated separately in literature. However non traditional optimization algorithms like Evolutionary Algorithms can be used to solve such problems simultaneously. The objective of this work is to develop solution techniques for multi-objective optimization of resource scheduling and time cost trade off problems simultaneously utilizing the power of non traditional optimization methods. Implementation will be done using principles of object oriented technology. Research Paper: A paper based on the planned work is being planned to be communicated to Journal of Construction Engineering and Management, ASCE.

RESEARCH PROJECTS
SITE LAYOUT PLANNING OF CONSTRUCTION SITES.
Guide : Dr. Arbind Kumar Singh, IIT Guwahati . Group Size : 2

Brief Summary
Site layout is an important aspect of construction planning, which potentially affects cost, time and safety in a project. The problem can be defined as allocation of predetermined temporary facilities like warehouses, job offices, workshops and batch plants to available sites so as to optimize an objective subject to layout constraints and requirements. The vector of decision variables is the location of all the $n_f$ facilities with the constraints being the uniqueness of location and site for a particular facility. The objective, typically is to minimize sum of product of resource flow (considered in a generic sense) between two facilities and distance between the locations of the two facilities. In this work a solution to the problem is attempted using ant colony optimization(ACO), an algorithm inspired by foraging behavior of real ants. Formulation of site layout problem in terms of a solution construction graph to be used by (artificial) ants has been attempted. Efficacy of the model has been demonstrated with a case study of a 12 facility problem. Research Paper: ``Site Layout of Temporary Construction Facilities using Ant Colony Optimization.'' accepted for presentation at ASCE 4th International Engineering and Construction Conference July 28, 2006,Los Angeles, US .

ANT COLONY OPTIMIZATION- ARTIFICIAL NEURAL NETWORK TECHNIQUE FOR BACK-CALCULATION OF PAVEMENT LAYER MODULI
Guide : Dr. M Amaranatha Reddy, IIT Guwahati. Group Size : 2


Brief Summary
Back-calculation of pavement layer moduli is the process of evaluating the layer moduli using pavement surface deflections measured by nondestructive testing equipment. A major problem faced by traditional back-calculation approaches is that the multi-dimensional surface represented by the objective function (related to the difference between calculated and measured surface deflections) may have many local minima and dependency on the initial seed moduli. Similarly non-traditional methods, though gives satisfactory results, requires huge computational time due to repeated calls to forward calculation program in the process of selecting best solution. To overcome these demerits, an attempt has been made in the present work, where Artificial Neural Network (ANN) and Ant Colony Optimization based program (ACO-ANNBACK) was developed. In this program, ANN technique was used to develop the forward calculation program to efficiently compute surface deflections using elastic layer theory and Ant Colony Optimization (ACO) meta-heuristic was used for back-calculation. The computational efficiency of ANNs combined with the robustness of the ACO gave highly satisfactory back-calculated layer moduli results (in terms of accuracy and very less computation time) for three and four-pavement layers Research Paper: The paper based on this work has been accepted for publication in 22nd Australian Road Research Board Conference, Canberra, Australia. A journal paper of this work is under preparation and soon will be communicated to International Journal of Road Materials and Pavement Design.

SUMMER PROJECT
DISCRETE OPTIMIZATION OF TRUSSES WITH ANT COLONY METAPHOR.
Guide : Self Guided Group Size : 3

Brief Summary
In structural design, cross sections must be chosen from a limited set of sections commercially available. Sizing optimization of trusses using ant colony optimization (ACO) was accomplished in this work. The objective of the this work is to develop a technique for optimal design of trusses allowing discrete variables alone, making use of the ant colony metaphor. Structural Analysis of trusses is done using FEM. A novel representation scheme for mapping the truss design to the minimum path length formulation is suggested. Stress deflection considerations are handled using constraints in the form of penalty functions. A number of examples are considered whereby the usefulness of the approach is demonstrated. The best design is obtained through a series of iterations. The results obtained through the present study in terms of the optimum volume of material are found to be intuitively optimal suggesting that the proposed method could be used for getting a good truss designs Research Paper: ``Discrete optimization of Trusses using ant colony metaphor'' presented at International Conference on World of Innovations in Structural Engineering, Hyderabad Dec.1-3,2004 and bagged the Best Paper Award.

SOFTWARE PROJECTS
CRITICAL PATH METHOD
Guide : Dr. Arbind Kumar Singh

Brief Summary
A software for critical path calculation in activity on edge networks (CPM) has been developed. Input for the software, is the duration of each activity and the start and finish event nodes. A linked list representation of the network graph has been used. Thereafter numbering of events is done using topological sorting algorithm. The algorithm has been suitably modified to calculate the early start time and latest start time of each node. Using these values the float for each activity is calculated and the critical path is identified. The time complexity of the algorithm is $\mathcal{O}(n + e)$, where $n$ is the number of activities and $e$ is the number of event nodes. The software was developed in the C programming language in Linux operating system.

BEAM DESIGN IN CONFORMANCE WITH INDIAN STANDARD CODE IS456:2000.
Guide : Dr. Anjan Dutta Group Size : 6

Brief Summary
Engineering design is an iterative process. Using a software, the designer can check and design economical structural sections quickly and reliably. A software for design for beam sections was prepared with this objective in mind. Guidelines from IS456:2000 have been used. Input required is the maximum positive and negative moments along the length of the beam, and the maximum shear force and torsional force, if any. T, L and rectangular beams can be designed and checked. A friendly graphical interface was prepared in Visual Basic to make the use easier. The design code was written in C.