Solving Job Shop Scheduling Problems with a Generic Bee Colony Optimization Framework

Job Shop Scheduling Problem (JSSP) is a Combinatorial Optimization Problem (COP) with NP-hard nature [ 14]. Hence, finding an optimum solution for a JSSP with large scale of data and constraints is complicated. This paper presents a generic Bee Colony Optimization (BCO) framework for solving JSSP. The proposed BCO framework realizes computationally the foraging behaviour and waggle dance performed by bees. It is also integrated with mechanisms such as elitism, local optimization and adaptive pruning. The framework is designed using the object-oriented approach where it contains a set of domain independent classes and abstract classes. The framework can be extended to solve other COPs and any enhancement added to the proposed framework will be applicable across all other COPs. A set of 82 JSSP benchmark instances were employed as the testbed and the results show that the proposed BCO framework is able to solve 54% of them to an average deviation percentage of <= 1% from known optimal or known upper bound.