Developing and Solving Multi Objective Bi-level Production-Routing problem models in deterministic and non-deterministic conditions

نوع: Type: thesis

مقطع: Segment: PHD

عنوان: Title: Developing and Solving Multi Objective Bi-level Production-Routing problem models in deterministic and non-deterministic conditions

ارائه دهنده: Provider: Farzane Adabi

اساتید راهنما: Supervisors: Dr. Amirsaman Kheirkhah

اساتید مشاور: Advisory Professors: Prof. Reza Tavakkoli-Moghaddam

اساتید ممتحن یا داور: Examining professors or referees: Dr. Parviz Fattahi, Dr. Donya Rahmani, Dr. parvaneh Samouei

زمان و تاریخ ارائه: Time and date of presentation: date 2022/06/28 time 8 am

مکان ارائه: Place of presentation: Amphitheater

چکیده: Abstract: The production-routing problem is a combination of the lot sizing problem and the vehicle routing problem. In this case, the manufacturer and distributor as the decision units, decide on the amount and time of production and also the distribution routes and amount of each vehicle. In the real world, the internal components of a system, such as manufacturer and distributor, can have more independent decision-making power and performance. This study, to bring the production-routing problem closer to real world problems, develops the production-routing problem in the aspects of uncertainty and multiplicity of decision makers. It is tried to purpose almost realistic models with more flexibility and usability, with the assumptions of a bi-level model and a range of multi-dimensional variations. These assumptions increase managers' decision-making areas and adapt to the real conditions of production and distribution chains. For this purpose, two main issues have been presented and developed in this research. In the first problem, the production-routing problem is designed with bi-levels of distributor and producer decision making, which in addition to paying attention to the presence of different decision makers, the production-routing problem has also been developed from the aspect of environmental protection and Possibility of outsourcing. Distributor, assumed as the leader and manufacturer as the follower. Due to the nature of being bi-level and multi-objective of this problem, the bi-level fuzzy goal programming approach has been developed as a precise solution. The analysis of numerical results of this part, compares the multi-objective bi-level production-routing problem with decision maker multi-objective production-routing problem by analyzing the sensitivity of outsourcing penalty changes. The results show that non-dominated solutions to a one-level problem are not stable and balanced answers to the problem in a competitive environment. By examining the usual indicators for multi-objective problems, the superiority of the rate of achievement simultaneously index in the multi-objective bi-level production-routing problem was obtained over the multi-objective production-routing problem. This clarified the nature of the behavior of the bi-level problem, so that the bi-level problem does not seek to find diverse and focused Pareto answers, but seeks to create a balance among decision-makers in order to achieve their goals. Continuing the analysis of numerical results examines the following: Sensitivity of target functions at high and low levels compared to each other and their tolerance limits, Sensitivity of outsourced goods and all retained goods to changes in outsourced fines, Fuzzy goal programming method parameter (tolerance of variable route and amount of goods sent). In the second problem, the production-routing problem is modeled with two levels of distributor and producer decision making, assuming demand uncertainty and shipping costs with a range of multi-dimensional variations. For this purpose, robust optimization and adjustable column-wise robust optimization approaches have been developed. To solve this problem, the active constraint Karush–Kuhn–Tucker algorithm with binary variables has been developed. In the following, the uncertainty in the demand parameter and the separate occurrence of this disturbance in the distributor objective function, inventory positive constraint, inventory capacity constraint and simultaneous occurrence of disturbance in these relationships, is analyzed. Sensitivity analysis of these uncertainties is presented in order to examine the variations in target functions of high and low level and outsourced goods in relation to changes in production costs and the ratio of demand uncertainty intervals. In the following, the effect of uncertainty in the transportation cost parameter is studied. The results of the uncertainty analysis make it clear that if the problem is simplified and a maximum of 30% disturbance is not assumed to occur in the demand parameter, the value of the distributor's objective function can be estimated up to 43 times less than reality. This value is a maximum of 25% for the manufacturer's objective function. Also, uncertainty in the cost of transportation with 100% disruption can be associated with a 12% increase in total costs. The meta-heuristic algorithm for optimizing the multi-objective particle swarm and the hybrid of multi-objective particle swarm optimization algorithm with Gams software are presented to solve the multi-objective bi-objective production-routing problem. After adjusting the parameters of the proposed algorithms, the performance of the algorithms is evaluated with the RAS rate of achievement simultaneously index. The results indicate the superior performance efficiency of the hybrid solution based on the simultaneous achievement rate criterion

فایل: ّFile: Download فایل