The diagram below summarises three major steps and equipment $($or 'machines'$)$ for the steel$-$making process of a Japanese company $($as presented in Zweben & Fox, Chapter $21).$

The converter refines pig iron into steel of required composition, the refinement equipment removes impurities from molten steel and$/$or adds alloy ingredients on molten steel to make high$-$grade steel, and a continuous caster casts molten steel into slabs, blooms or billets.

Assume that this production plant has three identical converters, nine different refining devices which are assigned according to the grade of the product, and five different continuous casters which are assigned according to the specification of the product. There is a fixed number of charges that should be scheduled for production in a day $($A charge is a work unit representing a fixed amount of molten steel$).$ FIRST$-$ Discuss two main reasons why operations research or other traditional analytical methods are not the most efficient for generating production schedules and for handling dynamic rescheduling requirements.

SECOND$-$Identify the potential intelligent agents for this system and specify their key attributes. Decide whether scheduling should be job$-$driven or machine$-$driven and state your assumptions