Hybrid meta-heuristic algorithm for solving multi-objective aggregate production planning in fuzzy environment

• Main Author: Kalaf, Kalaf, Bayda Atiya
• Format: Thesis
• Language: English
• Published: 2017
• Online Access: http://psasir.upm.edu.my/id/eprint/70930/1/FS%202017%2051%20IR.pdf; http://psasir.upm.edu.my/id/eprint/70930/

Aggregate production planning (APP) is considered as significant for efficient production systems. APP problems are considerably important in several manufacturing concerns. In actual APP problems, input data or parameter values, including resource, demand, cost, and objective functions, may be inaccurate. On the other hand, consideration of all parameters in an APP model makes the generation of a master production schedule deeply complicated especially in real-world APP problems, where input data or parameters are frequently imprecise (fuzzy) due to incomplete or un obtain able information and daily changes patterns of demand and manufacturers capacity (Sakalhet al., 2010). In addition, the APP problem based on the fuzzy environment becomes even more sophisticated as decision makers try to consider multi-objectives. Therefore, this study attempts to propose a novel scheme which is capable of dealing with these obstacles in APP problem. This schema takes into account uncertainty and makes a trade-off among conflicting multi-objectives at the same time. In addition, the proposed technique comprises of two main steps: first, some critical decisions about determining production rate and human resource planning (fuzzy data) are considered; next, decision about quantity and method of holding inventory and distribution of end product to customers was made. During the course of the present work, two fuzzy methods (modified Zimmermanns approach and modified angelovs approach ) and fourmeta-heuristics and hybrid meta heuristics including; simulated annealing (SA), modified simulated annealing (MSA), hybrid modified simulated annealing and simplex downhill (MSASD), hybrid modified simulated annealing and modified particle swarm optimization (MSAPSO) were proposed. For these proposed approaches, this study adopted a hybridization of a fuzzy programming, modify simulated annealing, and simplex downhill (SD) algorithm called Fuzzy-MSASD to resolve multiple objective linear programming APP problems in a fuzzy environment. The proposed strategy is dependent on modified Zimmermanns approach for handling all inexact operating costs, data capacities, and demand variables. The SD algorithm is employed to balance exploitation and exploration in MSA, thereby resulting in efficient and effective (speed and quality) solution for the APP model. Finally, the proposed approach was implemented in a real-world problem for Baghdad Soft Drinks Company to automate their APP. The findings showed that the proposed approach produced significant efficient solutions and achieved significantly low computational time for APP in large-scale problems.