Introduction (1) • A model by itself does not provide a solution to a problem, however, algorithms or solution techniques have to be developed to solve model • An algorithm is a step-by-step procedure that finds a solution to a model, and hence to the problem, in a finite number of steps • A number of algorithm have been proposed and these may be classified as: (a) Optimal algoritms and (b) Heuristic algorithms • All optimal algorithms developed for the layout problem require extremly high memory and computational time, and they increase exponentially as the problem size increases. Introduction (2) • The heuristic algorithms are devided into tree classes: 1. Construction algorithms: starting with an empty layout, they add one department (or a set of departments) after another until all the departments are included in the layout 2. Improvement algorithms; systematically modify the starting solution and evaluate the resulting modified solution. If it is better, the modification is made permanent. If not, the systematic modification is continued until it is n longer possible to produce better solutions 3. Hybrid (composite) algorithms; algorithms that use two or more types of solution techniques Algorithmic Approaches Layout algorithms can be classified according to: Type of input data; qualitative flow data – quantitative flow data Objective functions; minimizing of the sum of flows time distance – maximizing an adjacency score Format they use for layout representation; discrete – continuous representation Their primary function; layout improvement – layout construction Construction Algorithms • Construction algorithms generate a facility layout from scratch. • Starting with an empty layout, they add one department (or a set a departments) after another until all the departments are included in the layout • The main difference among the various construction algorithms relate to the criteria used to determine the: • First department to enter the layout • Subsequent department or departments added to the layout • Location of the first and subsequent departments in the layout • Example: Modified Spanning Tree Algorithm for Single-row Layout Problem, Graph Theoritic Approach CRAFT Computerized Relative Allocation of Facilities Technique (Armour, Buffa, and Vollman, 1963) Input data : from – to chart An improvement-type layout algorithm Departments represented in a discrete fashion It begins by determining the centroids of the departments in the initial layout, then calculates the rectilinear distance between pairs of department centroids and stores values in a distance matrix CRAFT next considers all-possible two-way (pairwise) or three-way department exchanges and identifies the best exchange ( maximal reduction in layout cost) CRAFT The next iteration starts with CRAFT once again identifying the best exchange by considering all- possible two-way or three-way exchanges in the (update) layout. The process continuous until no further reduction in layout cost can be obtained The final layout obtained in such a manner is also known as a two-opt (three-opt) layout CRAFT Initial CRAFT Layout Intermidiate CRAFT Layout (z = 2974 x 20 = 59,480 units) (z = 2953 x 20 = 59,060 units)
Final CRAFT Layout Final “massaged” layout obtained
(z = 2833.50 x 20 = 56,670 units) with CRAFT BLOCPLAN Departments arranged in bands which the number of bands is determined by the program and limited to two or three bands All the departments are rectangular in shape Input data: a relationship chart anda a from-to chart, the two charts can be used only one at a time when evaluating a layout Layout cost can be measured either by the distance-based objective or the adjacency-based objective BLOCPLAN uses the continuous representation BLOCPLAN may be used both as a construction algorithm and an improvement algorithm LOGIC Layout Optimization with Guillotine Induced Cuts (Tam, 1991) Input data: a from-to chart The layout is represented in a continuous fashion A construction and improvement algorithm LOGIC is based on dividing the building into smaller and smaller portions by executing successive “guillotine” cuts (straight lines that run from one end of the building to the other). Each cut is either a vertical cut or a horizontal cut LOGIC LOGIC LOGIC