Ajay Joneja

Automated MEP pipe routing faces inherent challenges in spatial coordination and multi-objective optimization. Traditional methods struggle with sequential spatial dependencies, where early-routed pipes often hinder system-level optimality. This study proposes a coarse-to-fine coordination framework for rapid design prototyping and high-quality layout generation. We introduce a header-to-header model treats connection positions as decision variables and employs Enhanced Conflict-Based Search (ECBS) for dynamic arbitration of routing sequences. Furthermore, a unified lifecycle cost model allowing the user to balance between setup and operation costs. Thus, the algorithm can generate a diverse solution set, allowing engineers to actively navigate the trade-off between construction economy and maintenance accessibility based on user-defined weights. Feasibility of our approach was validated through case studies. The framework can synthesize engineering-compliant 3D layouts in a few minutes. These results demonstrate its capability to significantly accelerate the design cycle by providing engineers with multiple valid design alternatives to facilitate early-stage planning.

The desire for sustainability and improved air quality has led architects to explore integrating vertical axis wind turbines (VAWT) in urban skyscrapers. However, the efficiency of such solutions is sensitive to the geometry of the wind channel. In this paper, we present a general technique for optimization of the wind channel geometry. Using parametric curves to define the profile of the channel, and by quantizing the location of the control points, we propose an experimental design approach to determine near-optimal channel geometry. The solution is further improved by interpolating the performance function so obtained via a statistical tool called kriging. The approach is tested by an experimental study, in which the parameters of the fluid dynamic model are determined by a series of wind tunnel tests.

This paper addresses the issue of placing a simple polygon (upon translation and rotation) on an isometric triangular grid such that the polygon contains the maximum number of triangles in its closure. This solves the problem left open in two recent papers titled “On the problem of the automated design of large-scale robot skin” and “An improved algorithm for the automated design of large scale robot skin” published in the IEEE Transactions on Automation Science and Engineering . Based on the properties of the grid, an improved algorithm is also presented. We also present some experimental results describing the use of this algorithm.

The problem of interpolating a quadrilateral free-form network with a quasi-developable patch is investigated in this paper. The boundaries of the quadrilateral are assumed to be degree 3 Bezier curves. The interpolated surface is a bi-cubic Gregory patch, which guarantees G¹ Continuity. All inner control points of the Gregory patch are determined by imposing G¹ continuity constraint and solving an optimization to maximize the surface developability. The optimization model minimizes the value of the integral Gaussian curvature over the entire patch. Examples are provided to illustrate the application of this approach in several industrial applications.