Approach for solving a scheduling problem for multipurpose batch chemical plants

An approach for solving a scheduling problem relevant to a multipurpose batch chemical plant with multiple production routes is proposed in the article. A mixed campaign operation mode is regarded. One mixed campaign may include both different routes of different products and more than one route for the single product. By selecting a suitable class of schedules, i.e. schedules involving consecutively carried out campaigns of compatible production routes and minimization of the schedule duration as a criteria for an estimation, a scheduling problem is formulated as a covering problem in terms of binary linear programming. An approximation method for solution of a scheduling problem which is intended to cope with two basic problems: (i) to avoid the large dimension of the initial task, and (ii) to obtain rapidly the optimal or near to optimal schedules of the desired class is presented. The first task of avoiding the high dimension of the initial problem is achieved by determination of the superstructure of the schedules of the desired class. The optimal or near to optimal schedules are determined by using the solution method based on the search tree applied to the superstructure obtained. The theoretical bounds of the desired schedules and the branching function are formulated. An algorithm is proposed to find the solutions. The procedure under consideration is illustrated by solving a simple example.