This project involved developing a sophisticated construction layout optimization application that addresses five critical objectives: hoisting time, risk, safety, hazard safety, and transportation cost. To solve this challenging multi-objective problem, I implemented and compared several cutting-edge metaheuristic algorithms, including:
The application provides an interactive web interface built with Svelte, delivering a seamless user experience and real-time feedback. The backend is powered by Golang, ensuring efficient computation and data management. For visualization, ECharts was integrated to display complex multi-dimensional Pareto fronts and objective trade-offs through intuitive charts and graphs.
This comprehensive tool enables construction managers and planners to make data-driven decisions by balancing multiple conflicting objectives, enhancing project safety and efficiency while minimizing costs.
In this project, I developed a full-stack invoice tracking system for monitoring the creation, modification, and deletion of invoices within a business accounting environment. The system was built with a Go (Golang) backend, providing a robust and efficient API for processing and storing invoice activity logs, and a responsive Vue.js frontend for user interaction and real-time updates.
Key features included activity logging, user role control, and detailed audit trails to ensure data integrity and transparency. The system also supported filtering, searching, and exporting logs, making it easier for accounting teams to trace changes and ensure compliance.
This project demonstrated my ability to build reliable, maintainable, and secure web applications with modern technologies and a focus on traceability and operational clarity.
The Fossa Optimization Algorithm (FOA) is a novel metaheuristic inspired by the hunting and social behaviors of the Fossa, a unique predatory mammal endemic to Madagascar. FOA operates through two distinct phases, each managed by a separate swarm, mirroring the Fossa's dual strategy of exploring new territories and exploiting known food sources.
In the Exploration Phase, a dedicated "Scout Swarm" simulates the Fossa's solitary and wide-ranging search for new prey and habitats. This phase emphasizes diversification and global search capabilities, allowing the algorithm to extensively explore the search space and avoid premature convergence to local optima.
Conversely, the Exploitation Phase is managed by a "Hunter Swarm" that models the Fossa's more focused and cooperative hunting tactics when a prey-rich area is identified. This swarm prioritizes intensification, refining solutions within promising regions of the search space to converge rapidly towards optimal or near-optimal solutions.
The seamless transition and interaction between these two swarms, mimicking the Fossa's adaptive foraging strategies, enable FOA to achieve a robust balance between exploration and exploitation, making it a promising candidate for solving complex optimization problems.
This project addresses the Resource-Constrained Project Scheduling Problem (RCPSP), a critical challenge in project management that involves optimally allocating limited resources to project activities over time. The primary objective is to develop and implement a novel metaheuristic approach to simultaneously maximize the Net Present Value (NPV) of the project and minimize its total completion time.
The proposed solution leverages the Multi-Objective Artificial Hummingbird Algorithm (MOAHA), a nature-inspired optimization algorithm, to efficiently explore the complex solution space of the RCPSP. MOAHA is designed to handle multiple conflicting objectives, providing a set of Pareto-optimal solutions that represent the best trade-offs between NPV maximization and project duration minimization.
The entire optimization framework, including the MOAHA implementation and its application to the RCPSP, is developed and simulated within the MATLAB environment. This project aims to demonstrate the effectiveness and robustness of the MOAHA in solving real-world project scheduling problems, offering a powerful tool for decision-makers to achieve superior project outcomes under resource constraints.
In this advanced layout planning problem, I optimized the placement of facilities within a defined boundary while simultaneously satisfying three critical objectives: minimizing risk, reducing hoisting time, and enhancing on-site safety. The solution was developed in MATLAB, using a hybrid multi-objective Artificial Hummingbird Algorithm (MO-AHA) enhanced with three distinct Opposite-Based Learning (OBL) strategies.
This hybridization significantly improved the algorithm's convergence behavior and solution diversity. The OBL mechanisms enabled the model to efficiently explore the search space by evaluating complementary solutions, while MO-AHA mimicked hummingbirds’ intelligent foraging behavior to balance the conflicting objectives.
The resulting Pareto front allowed stakeholders to select from a range of trade-off solutions tailored to operational priorities, such as speed versus safety or cost versus risk, providing a practical tool for decision-making in high-risk construction environments.
This project addressed the Capacitated Vehicle Routing Problem (CVRP) in the context of trash collection logistics in District 2, where the objective was to determine the most efficient routes for a fleet of waste collection vehicles with capacity constraints. I implemented the Osprey Optimization Algorithm (OOA)—a recent nature-inspired metaheuristic—in MATLAB to solve this complex, real-world combinatorial problem.
The algorithm intelligently assigned trash pickup points to vehicles, ensuring optimal route length while respecting vehicle capacity and operational limits. Through iterative refinement, OOA minimized the total distance traveled and balanced the load across the fleet, ultimately reducing fuel consumption and improving service reliability.
This work provided a scalable and adaptable solution for urban waste management, showcasing how advanced metaheuristics can drive real impact in smart city logistics.
In this phase, I focused on optimizing building design parameters to achieve energy-efficient temperature regulation. Using IESVE solely as a data generation tool, I simulated a range of building scenarios under different environmental and structural configurations. The resulting dataset was then used to train an Artificial Neural Network (ANN) in PyTorch, which served as a predictive model for indoor thermal performance.
To fine-tune the input parameters (e.g., insulation thickness, HVAC settings, material properties), I applied the Grey Wolf Optimizer (GWO)—a nature-inspired metaheuristic well-suited for continuous optimization tasks. The ANN-GWO hybrid enabled rapid evaluation of design alternatives without the need for repeated IESVE simulations, significantly speeding up the optimization process.
This approach demonstrated the power of surrogate modeling in building energy analysis, providing a scalable and intelligent solution for early-stage design decisions in sustainable architecture.
This component focused on the parametric design of glass installations to regulate natural light and indoor temperature, contributing to both energy efficiency and occupant comfort. I developed a custom Grasshopper plugin for Rhino, written in C#, which allowed designers to intuitively manipulate architectural parameters and immediately visualize their impact on environmental performance.
To optimize the design, I implemented the African Vulture Optimization Algorithm (AVOA)—a recent nature-inspired metaheuristic known for its strong balance between exploration and exploitation. The algorithm iteratively adjusted parameters such as glazing angle, tint level, and panel placement to maximize daylight utilization while minimizing thermal gain.
By combining computational design with intelligent optimization, this tool empowered architects and engineers to make data-driven decisions during the early stages of design, ensuring aesthetics and sustainability were seamlessly integrated.
In this phase of the project, I addressed the challenge of optimizing material placement within a set of predefined locations to reduce handling time and improve workflow efficiency. I developed an enhanced version of the Ant Lion Optimization (ALO) algorithm in Python, integrating two powerful strategies: mutation operators and opposite-based learning (OBL).
This hybridization significantly boosted the algorithm’s exploration capabilities and convergence reliability. The mutation component introduced diversity into the search space, helping to escape local optima, while opposite-based learning accelerated convergence by evaluating both current and opposite solutions during the search process.
The resulting system intelligently positioned materials based on frequency of use, access priority, and spatial constraints, delivering tangible improvements in layout efficiency and operational cost.
This project addresses the complex challenge of optimizing gravitational sewer system design, focusing on the strategic selection and configuration of manholes and pipes. The objective is to achieve a highly efficient and cost-effective sewer network that adheres to all necessary hydraulic and topographical constraints.
To tackle the multi-faceted nature of this optimization problem, a novel hybrid metaheuristic approach is proposed, combining the strengths of two powerful algorithms: the Salp Swarm Algorithm (SSA) and the Dragonfly Algorithm (DA). This hybridization aims to enhance the search capabilities, improve convergence speed, and avoid local optima, leading to more robust and globally optimal design solutions.
The SSA-DA hybrid algorithm will be employed to intelligently explore the vast design space, considering factors such as pipe diameters, slopes, manhole depths, and material costs. The project's outcome will be a refined methodology for designing gravitational sewer systems that minimizes construction expenses, operational costs, and environmental impact, thereby contributing to more sustainable and resilient urban infrastructure.
In this phase, I tackled the classic time–cost management problem in project scheduling by implementing a Moth–Flame Optimizer (MFO) enhanced with a Modified Adaptive Weight strategy in Python. The goal was to find optimal trade-offs between project duration and total cost under realistic constraints—such as resource limits and precedence relationships—by dynamically adjusting the MFO’s exploration–exploitation balance.
The modified adaptive weight approach recalibrates the influence of the “moth” population over iterations, improving convergence speed and solution diversity. This allowed the algorithm to more effectively navigate the non-linear time–cost surface, identifying Pareto-efficient schedules that reduce overall expenditure without unduly prolonging project timelines.
This component of the project focused on solving a facility layout optimization problem, where m facilities needed to be optimally placed across n predefined locations to minimize material handling costs and improve operational flow. I implemented the Ant Lion Optimization (ALO) algorithm in MATLAB to tackle this NP-hard combinatorial problem efficiently.
ALO, inspired by the hunting mechanism of antlions in nature, was used to intelligently explore the solution space and identify optimal or near-optimal layouts based on distance and flow matrices. The algorithm’s adaptive search capabilities allowed it to balance exploration and exploitation effectively, making it well-suited for the dynamic constraints of real-world facility planning.
By leveraging MATLAB for modeling, simulation, and result visualization, I created a solution that not only outperformed traditional heuristics but also offered a practical decision-support tool for layout planning in manufacturing or construction environments.
In this project, I developed a hybrid metaheuristic solution combining the Dragonfly Algorithm (DA) and Particle Swarm Optimization (PSO) to optimize demand ordering strategies over various Fixed Order Period (FOP) intervals — including 1-day, 3-day, 5-day, and 7-day cycles. Implemented in MATLAB, the algorithm was designed to balance responsiveness and cost-efficiency by dynamically adjusting order schedules based on demand patterns and operational constraints.
The hybrid approach leveraged the exploration strength of the Dragonfly Algorithm and the convergence speed of PSO, resulting in faster convergence to high-quality solutions. By modeling and minimizing total cost—including holding, shortage, and ordering costs—this system effectively identified optimal ordering frequencies for different time horizons.
This work demonstrated the potential of bio-inspired algorithms in solving complex, real-world inventory management and supply chain challenges with a high degree of accuracy and adaptability.
I'm always interested in new opportunities and collaborations. Feel free to reach out if you'd like to discuss a project or just say hello.