SCHEDULING IN FLEXIBLE MANUFACTURING SYSTEM (FMS) USING SOFT COMPUTING TECHNIQUES

Abstract

In today's highly competitive and fast-paced manufacturing industry, companies are increasingly turning to flexible manufacturing systems (FMS) to improve their efficiency and productivity. FMS is an automated manufacturing system that includes transport vehicles, automated storage, and a comprehensive computer control system, all working together to produce a wide variety of parts quickly and efficiently. FMS is a critical component of Industry 4.0, the fourth industrial revolution characterized by the integration of advanced technologies making it smart manufacturing process. Scheduling optimization is a crucial aspect of Flexible Manufacturing Systems (FMS) that involves determining the optimal sequence for producing multiple components and allocating the appropriate resources to each operation. The FMS scheduling optimization is of paramount importance for manufacturers, as it results in increased productivity and reduced production costs. By utilizing an efficient FMS scheduling optimization, manufacturers can achieve faster production times, higher throughput rates, and improved quality control. The optimization of FMS scheduling is a significant factor in the current Industry 4.0. The integration of advanced technologies with FMS scheduling optimization can lead to the development of smarter factories with improved efficiency, accuracy, and automation. As such, the optimization of FMS scheduling is a vital element in the success of modern manufacturing operations. Traditional optimization methods, such as linear programming and dynamic programming, have been used for scheduling optimization in manufacturing for several decades. However, these methods have limitations when it comes to solving complex scheduling problems in Flexible Manufacturing Systems (FMS), which are characterized by large vii search spaces, non-linear relationships, and combinatorial constraints. Metaheuristics, a class of optimization algorithms that use heuristic rules to explore the search space efficiently, have emerged as a powerful tool for solving complex FMS scheduling problems. Metaheuristic algorithms are inspired from natural phenomena and mimics it to find near-optimal solutions by iteratively exploring the search space, making probabilistic moves, and adapting to the search environment. These algorithms can handle multiple objectives, constraints, and uncertainty, making them suitable for FMS scheduling optimization. With the advancement of computing power and the availability of high performance computing platforms, metaheuristic algorithms have become even more useful in FMS scheduling optimization. In this research, three novel hybrid meta heuristic methods have been proposed: 1) GAPSOTS- An amalgamation of Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Tabu Search (TS) 2) HAdFA- The Hybrid Adaptive Firefly Algorithm and 3) HFPA- Hybrid Flower Pollination Algorithm. GAPSOTS is simple hybridization of classic meta heuristics without any adaptive features. The GAPSOTS suffered from local optima entrapment and convergence was impetuous. To address the premature convergence problem inherent in the classic Firefly Algorithm (FA), the researcher developed HAdFA that employs two novel adaptive strategies: employing an adaptive randomization parameter (α), which dynamically modifies at each step, and Gray relational analysis updates firefly at each step, thereby maintaining a balance between diversification and intensification. HFPA is inspired by the pollination strategy of flowers. Additionally, both HAdFA and HFPA are incorporated with a local search technique of enhanced simulated annealing to accelerate the algorithm and prevent local optima entrapment. viii The current study addresses FMS scheduling optimization for the following: • A Flexible Job Shop Scheduling Problem (FJSSP) was analysed, studied and tested with proposed meta-heuristics for several benchmark problems for multi-objectives of makespan (MSmax), maximal machine workload (WLmax), total workload (WLtotal), total idle time (Tidle ) and Total tardiness, i.e., lateness of jobs (Tlate ). • An FMS configuration, integrated with AGVs, Automatic storage and retrieval system (AS/RS) has been optimized using a Combined Objective Function (COF) with the aim of minimizing the machine idle time and the total penalty cost combinedly. In order to test the effectiveness of this optimization method, several problems were developed and tested by varying the number of jobs and machines for this particular FMS setup. • The concurrent scheduling of machines and AGVs in a multi-machine FMS setup for different layouts has been studied. This problem has been developed as a multi objective optimization with objectives to minimize the makespan, mean flow time, and mean machine idle time. Proposed meta-heuristics have been employed and tested on randomly generated example problems to evaluate their performance for this setup. These meta-heuristics have proven to be effective in finding optimal solutions, and their application can lead to improved efficiency and reduced costs in FMS setups. • Finally, a real-life case study was conducted in a Lube Oil Blending Plant, Faridabad, India. The proposed GAPSOTS and HAdFA are tested for three problems with varying jobs and machines for multi objectives. ix The corresponding computational experiments have been reported and analyzed. The suggested algorithms have been implemented and tested using Matlab R2019a, computing environment on an Intel Core™i7, with Windows 10. The results indicate that the proposed HAdFA tends to be more efficient among the proposed algorithms and consistently demonstrated to achieve not only optimal solutions but also new makespan values were found for some problems. The efficiency of HAdFA can be attributed to the adaptive parameters integrated into it. This algorithm significantly improves convergence speed and enables the exploration of a large number of rich optimal solutions.