4. GA OPTIMISATION METHOD
4.2. GA as SC Optimisation Tool
The majority of supply chain approaches relates to optimisation of ordering and stock levels and thereby efficient supply chain management. The genetic algorithm methodology has been used with different approaches, namely on supply chain optimisation and to determine the best business practices.
Truong & Azadivar (2003) combine simulation, mixed integer programming and a genetic algorithm. The genetic algorithm provides a mechanism to optimise qualitative and policy variables, and the mixed integer programming model reduced computing efforts by manipulating quantitative variables.
Finally simulation is used to evaluate performance of each supply chain configuration with non-linear, complex relationships and under more realistic assumptions.
Vergaraet al. (2002) developed an evolutionary algorithm (EA) for optimal synchronisation of supply chains, using the economic delivery and scheduling model and analyse supply chains dealing with multiple-components. The EA is shown to be much faster at solving large problems than an enumeration procedure and exhibits robust behaviour when tested on a variety of different problem parameters.
Optimisation is the methodology for improving the quality and desirability of a product or product concept.
Jeonga et al. (2002), for building a generic forecasting model applicable to SCM, proposed a linear causal forecasting model and its coefficients determined using the proposed genetic algorithms (GA), canonical GA and guided GA (GGA). Gonsalves et al. (2007) used a Genetic Algorithm for the operational optimisation of collaborative systems, where the cost function to be optimised was to find the minimum value of the cost function under operational constraints. Senouci & El-Rayes (2009) presented a model with multi-objective optimisation that provides new and unique capabilities including generating and evaluating optimal/ near-optimal construction resource utilisation and scheduling plans that simultaneously minimise the time and maximise the profit of construction projects. Sourirajana et al. (2009) explored the use of GAs to solve the Single Product Network Design Model with Lead time and Safety Stock Considerations.
O’Donnell et al. (2006) employed a genetic algorithm (GA) to reduce the bullwhip effect and cost in the MIT beer distribution game. The GA is used to determine the optimal ordering policy for members of the SC. The paper shows that the GA can reduce the bullwhip effect when facing deterministic and random customer demand combined with deterministic and random lead-times. This paper examined the effect of sales promotion on the ordering policies and shows that the bullwhip effect can be reduced, even when sales promotions occur in the SC. The same authors, (O’Donnell et al., 2009) studied the detrimental effect of promotions on the supply chain, one of the main causes of the bullwhip effect, and a genetic algorithm was proposed to reduce these negative effects. GAs were used to dampen the impact of the bullwhip effect and can be used to assist supply managers in predicting reorder quantities along the supply chain.
Daniel & Rajendran (2005) studied the performance of a single-product serial supply chain operating with a base-stock policy and to optimise the stock levels in the supply chain so as to minimise the total supply chain cost (TSCC), comprising holding and shortage costs at all the installations in the supply chain. The effectiveness of the proposed GA (in terms of generating base-stock levels with minimum TSCC) is compared with that of a random search procedure. In addition, optimal base-stock levels are obtained through complete enumeration of the solution space and compared with those yielded by the GA. It is
found that the solutions generated by the proposed GA do not significantly differ from the optimal solution obtained through complete enumeration for different supply chain settings, thereby showing the effectiveness of the proposed GA.
Lu et al.’s (2009) research presents an extension to the genetic algorithm approach to reducing the bullwhip effect by investigating the individual efficient or responsive strategy for each member in different online supply chains. Four types of supply chain structure, by positioning the decoupling point, were investigated to determine if the genetic algorithm (GA) can help find optimal ordering policy and lead time for each member and, at the same time, reduce the impact of the Bullwhip effect and total mean cost across the online supply chain. They showed that the optimal supply chain structure that presents better performance on both the total lead time and the mean cost should be employed.
Vijayalakshmi et al. (2011) attempt to design an Intelligent Forecasting Engine which uses a combination of forecasting techniques. The design was based on the use of Genetic Algorithms, for selecting the best methods to combine for forecasting. Radhakrishthinan et al. (2010 – 2 articles) wrote several articles where GA was used for minimising the total supply chain cost through the reduction of holding and shortage cost in the entire supply chain: stock optimisation in the supply chain is distinctively determined to achieve minimum total supply chain cost. Azadeh et al. (2010) address the successful application of GA-simulation to simulation model optimisation and design, through the stochastic behaviour of their supply chain system. Zhu et al. (2011) used a genetic algorithm is to solve the supplier’s replenishment model, where experiment results demonstrated the feasibility and the effectiveness of the replenishment strategy. Priya & Iyakutti (2011) presented an approach to optimise the reorder level (ROL) in the manufacturing unit taking consideration of the stock levels at the factory and the distribution centres of the supply chain, which in turn helps the production unit to optimise the production level and minimising the stock holding cost. A genetic algorithm is used for the optimisation in a multi-product, multi-level supply chain in a web enabled environment: the prediction of optimal ROL enables the manufacturing unit to overcome the excess/ shortage of stock levels in the upcoming period.
The main objective of this section is to highlight the potential of Genetic Algorithms as an optimisation tool for different types of supply chain realities. The field of application of this methodology is immense as described previously: optimise the production levels, minimise stock holding costs, optimise supplier’s replenishment and ordering policies, and optimise forecasting methods, optimise synchronisation of supply chains, optimal scheduling plans, reduce these negative effects of marketing promotions, just to name a few. From the literature review the article of greater interest is of O’Donnell et al. (2006) who proved that by using historical data, the optimal ordering policy for an SC could be found by employing GAs. According to these authors, by employing the ordering policies determined to be optimal, the bullwhip effect in SCs will be reduced.
In this work a genetic algorithm (GA) is proposed to determine the equation parameters that optimise the ordering levels, and minimising the holding and shortage costs in the entire supply chain, maximising overall profit. Simulation will then be used to evaluate the base-stock levels and the ordering parameters generated by the GA. The proposed GA is evaluated with the consideration of a variety of supply chain settings in order to test for its robustness of performance across different supply chain scenarios. Both the optimisation and simulations are performed in Java and the output data is then analysed in Excel.