Keywords: artificial life, immune algorithm, optimal allocation, shopping mall simulation, underground mall.

This paper aims to develop a method of providing the optimal allocation of an underground shopping mall by using Immune Algorithm and simulation. Stores in a shopping mall are, in general, grouped and allocated at some section, in order that customers can find out the suitable shops among various shops and compare them with ease. However, there arises a question that such an allocation of stores is optimal or efficient for the underground mall. In this study, an attempt is made to investigate the optimal allocation from the standpoint of attracting customers as many as possible, by using Immune Algorithm (IA) and Artificial-life (A-life) simulation.

IA is employed for the optimization process, whereas A-life simulation is used in the evaluation of the superiority of the alternatives. IA is one of optimization methods, which simulates an antibodies and self-regulated mechanism of the immune system possessed in living beings in an engineering manner. IA has such an advantage that it can produce multiple optimal meaningful solutions with different characteristics. A-life was invented to realize living things regardless if they are hard or wet. A numeric application example is presented to demonstrate the efficiency of the proposed method, in which the East Mall of underground mall in Osaka city is employed. Stores in the underground shopping mall in Osaka are roughly divided into four groups; restaurants, fast food shops, grocery shops, and service stores. A-life is applied to the evaluation in IA. The evaluation is done by counting the number of the artificial living things visiting the store. The artificial living things are allocated on the virtual map of the underground mall. They move a certain number of times. This system can provide two or more optimal allocations of stores without changing the kind of and the number of stores in the simulation.

The parameters of an artificial living thing consist of view, direction, interesting degree to stores, time, and energy. Artificial living things affect each other according to their existence density. An artificial living thing recognizes other artificial living things by the view. Then, the artificial living thing calculates the surrounding density and moves to the place where there is coarse density or condensed density. It is summarized that an artificial living thing recognizes the stores influenced by the surrounding density of other living things, gets interested in the stores due to the interesting degrees, and enters into some store by the prescribed probability. The parameters of IA are set as individual number=100, memory cell=20, the number of generation=300, gene length=66. The number of artificial living things generated for the initial population and the time steps of movement are changed to examine their influences. Furthermore, the influence of map size is examined by several bitmaps.

In this study, more effective allocations than the present allocation of East Mall could be found out. Considering the optimal allocation and semi-optimal allocations, it is concluded that the number of customers becomes larger, when the stores are allocated with some scatter and balanced grouping. However, general rules relating optimal and semi-optimal allocations could not be discovered.

The future direction of this study will be to apply a larger map data as Osaka underground mall and to develop a more definite process to decide the desire of people.

1

Koki Ishida, "Immune System and The Application", Corona Co, Japan, 1998

2

Riku Hattori, "World of Artificial Life", Ohmu publishing, 1994

3

Jhon.J.Fruin, "Space of Pedestrian", Kashima laboratory publication meeting, 1974

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £85 +P&P)