Multi-objective scheduling in an agent based Holonic manufacturing system

Abstract

The present paper is aimed at multi-objective scheduling in an agent based holonic manufacturing system to satisfy the goal of several communities namely the product, the resource, and the organization simultaneously. In this attempt, first a multi criteria based priority rule is developed following Simple Additive Weight (SAW) method under Multi Criteria Decision Making (MCDM) environment to rank the products. Accordingly, the products are allowed to select a particular resource for execution by negotiation considering minimum time as criterion. The interests of different communities are accomplished by allocating the ordered rank of products to the ordered rank of resources. Conflict, if arises between products and resources, are resolved by introducing the concept of Early Finish Time (EFT) as criterion for task allocation. A scheduling algorithm is proposed for execution of the rule. In view of machine failure, a cooperation strategy is evolved that also optimizes reallocation of the incomplete task. It is concluded that the proposed scheduling algorithm together with the disturbance handling algorithm are poised to satisfy the agent's local objective as well as organization's global objective concurrently and are commensurable with multi agent paradigm.

Handling machine breakdown for dynamic scheduling by a colony of cognitive agents in a holonic manufacturing framework

“…The negotiation based scheduling procedure remains same as reported in our earlier work (Jana et al 2014) under normal situation. The products get the opportunity for negotiation with the resources according to the rank, settle the agreement by negotiation, and prepare the local schedules.…”

“…The work is essentially an augmentation of our earlier work (Jana et al, 2014) in regard to achieving improved performance under disturbance by cognitive behavior of the agents. The objective is to justify the credentials that cognitive agents perform better than the ordinary agents under uncertainty.…”

“…To investigate the performance of cognitive agents in comparison to the ordinary agents, we consider the same platform (Jana et al, 2014) having a resource holon comprising three resources (Turning centers) (R1 to R3) and a product holon having eight products (P1 to P8). Additionally, the times of failure of the resources are also considered unchanged.…”