Evaluating the performance of elevator group control algorithms using a three-element new paradigm

Abstract

This paper presents a new paradigm for assessing the effectiveness of up-peak elevator group control algorithms. The new paradigm can be very effective in providing a mechanism for objectively assessing and comparing elevator group control algorithms. It is built around three essential components: idealised optimal benchmarks; random scenario testing; and progressive introduction of reality. An idealised optimal benchmark is the starting point for calculating an analytical upper bound for the performance of any algorithm. It provides a reference for comparing the performance of all algorithms. Random scenario testing is used to subject the elevator group controller to a randomly generated scenario (usually of passenger origin-destination pairs). The response of the group controller to a randomly generated scenario is recorded, and more scenarios are generated and added. The overall response (e.g., average of all responses) of the group control algorithm to the large number of scenarios represents an objective measure of its efficacy. The random scenario testing is first carried out under idealised or partially idealised conditions. Under the third component of the new paradigm, the conditions are gradually made more realistic and better reflective of reality. This third element is called the progressive introduction of reality. Practical application: This paper presents to the designer of the elevator group controller a new paradigm for assessing the benchmark against which he/she is working. The designer can be confident that whatever up-peak group control algorithm he/she develops, it cannot exceed this upper benchmark. This has important practical applications in benchmarking the performance of up-peak group controllers. It can also be used by consultants and clients to mediate elevator group controller performance claims presented by elevator manufacturers.

Uncertainty-Aware Robustness Assessment of Industrial Elevator Systems

“…Correspondingly, all uncertain situations caused by single uncertain factor and t-way interactions (see Definition 7) among several uncertain factors in uncertain factor set 𝑈 can be represented by set 𝑈 𝑆 = {𝑢𝑠 𝑜 |𝑜 = 1, 2, ..., 𝑁 𝑆 }, where 𝑢𝑠 𝑜 is the 𝑜th uncertain situation caused by 𝑈 , 𝑁 𝑆 is the number of uncertain situations, there are 𝑁 𝑆 = 𝑢𝑛 𝑡 =1 𝐶 𝑡 𝑢𝑛 = 2 𝑢𝑛 − 1 uncertain situations in total, which can be simulated by 2 𝑢𝑛 − 1 types of uncertain traffic profiles (see Definition 6). For example, to investigate uncertain factor set 𝑈 with three uncertain factors {𝑢 1 = Mass (M), 𝑢 2 = Loading Time (L), 𝑢 3 = Unloading Time (U )}, the total number of uncertain situations caused by 𝑈 is 7, of which 3 (𝐶 1 3 ) are caused by single uncertain factor {𝑢𝑠 1 = 𝑢𝑠𝑀, 𝑢𝑠 2 = 𝑢𝑠𝐿, 𝑢𝑠 3 = 𝑢𝑠𝑈 }, 3 (𝐶 2 3 ) are caused by 2-way interactions {𝑢𝑠 4 = 𝑢𝑠𝑀 − 𝐿, 𝑢𝑠 5 = 𝑢𝑠𝑀 − 𝑈 , 𝑢𝑠 6 = 𝑢𝑠𝐿 − 𝑈 } and 1 is (𝐶 3 3 ) caused by 3-way interactions {𝑢𝑠 7 = 𝑢𝑠𝑀 − 𝐿 − 𝑈 }. Studying interactions among uncertain factors is important since effects (e.g., degraded QoS) might only be visible when a dispatcher faces uncertainties from a less number of uncertain factors simultaneously.…”

“…They increased the current from 1.5A to 9.9A and observed that the maximum air-cooling capacity can be obtained at 9.4A. Al-Sharif et al [3] proposed a paradigm with three components for testing the performance of elevator group control algorithms using idealized benchmarks, random scenario testing, and gradually making scenarios more realistic. Wang et al [51] tested the seismic performance of a functional traction elevator in a reinforced concrete building under earthquake.…”

Comprehensive analysis of elevator static sectoring control systems using Monte Carlo simulation

“…Due to the reduction in the round trip time caused by the effects in 2 and 3 above, a reduction in the number of passengers boarding the car results (leading to smaller car loading); or more passengers can be taken into the elevator car (leading to an increase in the handling capacity). Algorithms that attempt to minimize the number of stops and the highest reversal floor are described in Al-Sharif et al, 24,25 and Al-Shrarif. 26…”